chemical sensing employing ph sensitive emeraldine …
TRANSCRIPT
CHEMICAL SENSING EMPLOYING pH SENSITIVE EMERALDINE BASE THIN
FILM FOR CARBON DIOXIDE DETECTION
Except where reference is made to the work of others, the work described in this dissertation is my own or was done in collaboration with my advisory committee. This dissertation does not include proprietary or classified information.
________________________________________ Mihai Irimia-Vladu
Certificate of Approval: ______________________________ ______________________________ Aleksandr L. Simonian Jeffrey W. Fergus, Chair Associate Professor Associate Professor Mechanical Engineering Department Mechanical Engineering Department Auburn University Auburn University ______________________________ ______________________________ ZhongYang Cheng Curtis G. Shannon Assistant Professor Professor Mechanical Engineering Department Chemistry Department Auburn University Auburn University
______________________________ Stephen L. McFarland Acting Dean Graduate School
CHEMICAL SENSING EMPLOYING pH SENSITIVE EMERALDINE BASE THIN
FILM FOR CARBON DIOXIDE DETECTION
Mihai Irimia-Vladu
A Dissertation
Submitted to
the Graduate Faculty of
Auburn University
in Partial Fulfillment of the
Requirements for the
Degree of
Doctor of Philosophy
Auburn, Alabama August 7, 2006
iii
CHEMICAL SENSING EMPLOYING pH SENSITIVE EMERALDINE BASE THIN
FILM FOR CARBON DIOXIDE DETECTION
Mihai Irimia-Vladu
Permission is granted to Auburn University to make copies of this dissertation at its discretion, upon request of individuals or institutions and at their expense.
The author reserves all publication rights.
_____________________________ Signature of Author
______________________________ Date
iv
DISSERTATION ABSTRACT
CHEMICAL SENSING EMPLOYING pH SENSITIVE EMERALDINE BASE THIN
FILM FOR CARBON DIOXIDE DETECTION
Mihai Irimia-Vladu
Doctor of Philosophy, August 7, 2006 (B.S. University of Craiova, Romania, 1997)
194 Typed Pages
Directed by Jeffrey W. Fergus
Respiration, or CO2 evolution, is a universal indicator for all the biological
activities. Among many potential applications, the measurement of CO2 evolution has
been found to be a rapid and nondestructive means for examining microbial
contamination of food.
The sensor developed in this work consists of a thin emeraldine base-polyaniline
(EB-PAni) film. In the first half of the project the effect of carbon dioxide over the
conductivity of a composite film of emeraldine base polyaniline and poly(vinyl alcohol)
in N-methyl pyrrolidone (NMP) respectively was tested. Argon gas or mixture of argon
and 5% CO2 were circulated through the glass cell containing the polymer film deposited
on interdigitated electrode and exposed to specific humidity levels fixed by aqueous
supersaturated salt solutions. In the second half of the project, a thin emeraldine base film
v
in NMP was directly deposited on interdigitated electrode and the respective sensor
inserted in water. Carbonic acid solutions of various pHs were generated by bubbling
specific mixtures of carbon dioxide and argon. Conductivity measurements were
performed by impedance spectroscopy throughout the project. The sensing mechanism is
based on intermediate stages of the transformation of the emeraldine base polyaniline to a
conductive salt type (ES-PAni). This EB-ES transformation is the consequence of the
exposure of EB-PAni to a protonic acid and is accompanied by a change in the
conductivity of the polymer film. Carbonic acid, unfortunately, is a very weak acid and is
unable to induce a conductivity change, but the intermediate steps that predetermine this
transformation are detected by impedance spectroscopy even when the overall
conductivity of the film is unchanged.
The composite thin film developed in the first part of the project showed poor
sensing characteristics: limited dynamic range, drift, instability and slow time response.
However, the sensor design employed in the second half of this work, coupled with
impedance spectroscopy measurements, revealed valuable information about conduction
mechanisms at pH levels were the overall conductivity of the film remained unchanged.
Typical impedance spectra for the emeraldine thin films for a frequency sweep
between 3.2 E7 to 1 Hz shows a single semicircle. The overall conductivity of the film
(5×10-4 S/cm) does not change when CO2 is bubbled through the water in which the
sensor is immersed, but an additional semicircle starts to appear at low (less than 200 Hz)
frequency corresponding to lowering the pH of the solution below 5.0. The original
semicircle diminishes in size but maintains its initial peak frequency. The EB film is very
sensitive to pH changes, therefore an additional semicircle appears in unpurified argon
vi
gas due to the reduction of the pH of water solution to 4.65. The same mechanism is
displayed in hydrochloric acid solutions of various pH. The formation of the second
semicircle depends on the initial conductivity of the emeraldine base film, a film
displaying an initial conductivity of 4.8 × 10-3 S/cm forming the second semicircle at a
pH of 5.85. The appearance of the second semicircle is most likely due to a preferential
protonation in the insulating matrix of the polymer film. The overall conductivity of the
film increases when the level of protonation in the insulating portion of the film reached a
level close to the protonation level in the scattered metallic islands, allowing the electron-
hopping mechanism to became active.
The sensor output is stable and reproducible even after 11 months passed from the
polymer film deposition.
vii
ACKNOWLEDGEMENTS
The author expresses his gratefulness to Dr. Jeffrey W. Fergus for all his
academic guidance, help and encouragement for the pursuit of this degree.
The author also likes to convey thanks to Dr. Zhong. Y. Cheng, Dr. Aleksandr
Simonian, Dr. Curtis Shannon and Dr. Eric Bakker for many insightful suggestions and
kind help.
Acknowledgements are due to Mr. Roy Howard and Ms. Shirley Lyles for their
unconditional support and patience, and to all the friends and colleagues who assisted the
author in various ways during his years of studies.
Most importantly, the author pays tribute to his wife, Marina for her
encouragement and continuous support throughout his education, to his son, Cristian for
bringing happiness in his family life, and to his parents, grandparents and in-laws for help
and continuous support.
Finally, the author would like to acknowledge the financial support of this work
from the Materials Research and Education Center.
viii
Style manual or journal used: Synthetic Metals Computer Software used:
Microsoft Word, Microsoft Excel, DataThief
ix
TABLE OF CONTENTS
LIST OF FIGURES............................................................................................................xi
LIST OF TABLES ...........................................................................................................xiv
1. INTRODUCTION AND RESEARCH OBJECTIVES................................... 1
1.1. Safety Issues Faced in Food Chain Industry.................................................... 1
1.2. Carbon Dioxide Evolution Rate Measurement................................................ 2
1.3. Classification of CO2 Sensors.......................................................................... 3
1.3.1. Electrochemical Techniques................................................................ 3
1.3.1.1. Potentiometric Type CO2 Sensors: Description .............................. 3
1.3.1.2. Amperometric Type CO2 Sensors ................................................. 16
1.3.1.2.1. Reducing Specie; Electrode Material ............................................. 18 1.3.1.2.2. Gas Permeable Membrane ............................................................. 20 1.3.1.2.3. Electron Transfer Studies ............................................................... 24 1.3.1.2.4. Reduction of O2 and CO2 at Micro-electrodes ............................... 25 1.3.1.2.5. Characteristic Values Reported...................................................... 26 1.3.1.2.6. Conclusions .................................................................................... 27
1.3.1.3. Conductometric Type CO2 Sensors ............................................... 27
1.3.1.4. Coulometric Type CO2 Sensors..................................................... 30
1.3.2. Spectroscopic Techniques ................................................................. 34
1.3.3. Other Methods of Detection .............................................................. 37
1.4. Conductive Polymers as Carbon Dioxide Sensors ........................................ 39
1.4.1. Conductive Polymers: Description .................................................... 39
1.4.2. Insulator to Metal Transition Process ................................................ 40
1.4.3. Polyaniline: Description .................................................................... 43
1.5. Impedance Spectroscopy: Description .......................................................... 58
1.6. Project Objective ........................................................................................... 63
2. EXPERIMENTAL SECTION....................................................................... 64
2.1. Suspended Sensor .......................................................................................... 64
2.1.1. Conductive Polymer Preparation....................................................... 64
2.1.2. Polymer Film Deposition and Interrogation ...................................... 66
2.1.3. Experimental Setup............................................................................ 67
x
2.1.4. Experimental Conditions for Material Analysis ................................ 69
2.2. Immersed Sensor ........................................................................................... 72
2.2.1. Conductive Polymer Preparation....................................................... 72
2.2.2. Polymer Film Deposition and Interrogation ...................................... 72
2.2.3. Experimental Setup............................................................................ 73
2.2.4. Measurement Conditions ................................................................... 74
3. RESULTS AND DISCUSSION.................................................................... 77
3.1. Suspended Sensor .......................................................................................... 77
3.1.1. Experimental Results ......................................................................... 77
3.1.2. Sensor Performance Evaluation......................................................... 86
3.1.3. Comparative Results .......................................................................... 87
3.1.4. Materials Characterization................................................................. 89
3.1.4.1. FTIR Spectroscopy Characterization............................................. 89
3.1.4.2. Differential Scanning Calorimetry Characterization ..................... 95
3.1.4.3. X-Ray Diffraction Characterization .............................................. 96
3.1.4.4. UV-Visible Spectrophotometry Characterization........................ 100
3.1.4.5. Materials Characterization Discussion ........................................ 101
3.1.5. Possible Explanations for Poor Performance .................................. 103
3.1.5.1. Tautomerism ................................................................................ 103
3.1.5.2. Limited Protonation in Emeraldine Base..................................... 105
3.1.5.3. Limited Amount of Water in PVA Matrix .................................. 107
3.2. Immersed Sensor ......................................................................................... 109
3.2.1. Experimental Results ....................................................................... 109
3.2.1.1. Conductivity Response in Hydrochloric Acid Solutions............. 111
3.2.1.2. Conductivity Response in Carbonic Acid Solutions ................... 122
3.2.2. Complexation Effect Study.............................................................. 126
3.2.3. Response of the Sensor in Ar-CO2 mixtures ................................... 132
3.2.4. Conclusions...................................................................................... 138
4. FUTURE WORK......................................................................................... 140
5. REFERENCES ............................................................................................ 142
6. APPENDIX…………………………………………………………….......162
xi
LIST OF FIGURES Figure 1. First generation Severinghaus type carbon dioxide sensor.................................. 5 Figure 2. Miniaturized version of the Severinghaus-type CO2 sensor employing liquid ion exchange membrane. ........................................................................... 9 Figure 3. Iridium oxide tip miniaturized Severinghaus type CO2 sensor.......................... 14 Figure 4. Sensing concept of the amperometric CO2 sensor............................................. 18 Figure 5. Schematic of the electrical diagram for ISFET. ................................................ 30 Figure 6. Schematic of the coulometric carbon dioxide sensor. ....................................... 31 Figure 7. Oxidation of a fully reduced aniline repeat unit. ............................................... 44 Figure 8. Polyaniline in various oxidation states. ............................................................. 46 Figure 9. The resonance-stabilized form of the emeraldine salt polymer. ........................ 48 Figure 10. Protonation induced charge unpairing in polyaniline. The counterion display is omitted from the schematic.............................................................. 50 Figure 11. Self doping of emeraldine base........................................................................ 51 Figure 12. a) Schematic representation of the acid-base and redox reactions between different forms of polyaniline; b) Thermodynamical stability domains of the pH for the different forms of polyaniline. ................................................. 56 Figure 13. Impedance plotted as a planar vector using rectangular and polar coordinates. ...................................................................................................... 59 Figure 14. The sketch of the interdigitated electrode........................................................ 66 Figure 15. The setup utilized in conductivity measurements for suspended sensor. ........ 67 Figure 16. The equivalent circuit schematic associated with the measured data in the suspended sensor case; This circuit will also be employed in the immersed sensor case....................................................................................................... 69 Figure 17. Experimental setup for the conductivity measurements in the immersed sensor case....................................................................................................... 73 Figure 18. Equivalent circuit schematic for the immersed sensor case. ........................... 74 Figure 19. The CO2 detection mechanism. ....................................................................... 77 Figure 20. Impedance spectra recorded by Impedance Gain Phase Analyzer .................. 78 Figure 21. Resistance of the composite film in Ar and Ar + 5% CO2 in a fixed humidity level of RH ~ 30% (cycle 1 in Table 4). ........................................... 80 Figure 22. Trend of the resistance response of the composite film in Ar and Ar + 5% CO2 at a fixed humidity level of ~ 30%. ........................................... 81 Figure 23. Logarithm of resistance versus relative humidity in argon and Ar + 5% CO2
for the composite polymeric film. The film was tested more than once in both magnesium and sodium chloride respectively. Each particular cycle was circled for the ease of visualization........................................................... 82 Figure 24. Resistance difference between Ar and Ar + 5% CO2 (logarithmic scale). ...... 83 Figure 25. Response time (t90) reached by the composite film for ~60% RH (cycle 2 in Table 4)........................................................................................... 84
xii
Figure 26. Comparison between FTIR spectra of emeraldine base without heat treatment and heat treated at various temperatures for 1 hour in helium atmosphere. ..................................................................................................... 90 Figure 27. Schematic of crosslinking process. .................................................................. 92 Figure 28. DSC thermogram of emeraldine base held isothermal for 30 minutes at 100ºC and ramped up 20ºC/min to 400ºC. ................................................... 96 Figure 29. X-ray diffraction of emeraldine base treated at various temperatures............. 97 Figure 30. UV-Visible spectrum of emeraldine base samples treated for one hour at various temperatures in helium atmosphere. ............................................. 101 Figure 31. Schematic of the tautomerism process........................................................... 105 Figure 32. Doping percentage as a function of pH of protonating solution for the class I of emeraldine base............................................................................... 106 Figure 33. Percent water absorbed in poly (vinyl alcohol)-NMP plasticized film, at various humidity levels. ............................................................................. 108 Figure 34. Impedance spectra in pure water and water where pure CO2 was bubbled. .. 110 Figure 35. Total resistance of emeraldine base thin film vs. pH in aqueous HCl solutions.................................................................................................. 111 Figure 36. Comparison of the conductivity responses as a function of pH for the two classes of emeraldine base: the EB-I reported by MacDiarmid [161] and the EB-II employed in the current project. ............................................. 113 Figure 37. Comparison of the doping percentage vs. pH responses for the two classes of emeraldine base reported in the literature: MacDiarmid [161] and Jozefowicz [234]; The data points shown on the graph belong to MacDiarmid, being reproduced by DataThief software close to the one originally published............................................................. 114 Figure 38. The formation and evolution of the second semicircle with variation of pH level. .............................................................................................................. 115 Figure 39. Appearance of the second semicircle as a horizontal line at pH=5.03 and its subsequent evolution with decreasing pH level. ...................................... 115 Figure 40. Resistance variation of the NMP-plasticized emeraldine base thin film with time at different pH levels, showing the onset of conductivity change at pH = 2.25. ...................................................................................... 116 Figure 41. Titration curve of hydrochloric acid in water and resistance response (Rhigh frequency) of the emeraldine film as a function of hydrochloric acid concentration. ................................................................................................ 117 Figure 42. Capacitance as a function of pH associated with the sensor response in hydrochloric acid........................................................................................... 118 Figure 43. Capacitance vs. square root of time for the emeraldine base thin film equilibrated for various times in pH = 1.75 ................................................... 119 Figure 44: Capacitance vs. square root of time for the emeraldine base thin film equilibrated for various times in pH = 1.11 ................................................... 120 Figure 45. Impedance as a function of pH of hydrochloric acid solution for two fixed frequencies. ........................................................................................... 121 Figure 46. Sensor resistance in alternating cycles involving alternating bubbling of CO2, purified and unpurified Ar through water. Graph labels:
xiii
“H2O”- pure water exposed to room atmosphere air, “CO2”, “Ar purified” and “Ar-unpurified”- bubbling of respective gas in pure water. The equivalent circuit model gives the data points on graph. ............................... 123 Figure 47. Sensor resistance in alternating cycles involving alternating bubbling of CO2, purified and unpurified Ar through water. Graph labels: “H2O”- pure water exposed to room atmosphere air, “CO2”, “Ar purified” and “Ar-unpurified”- bubbling of respective gas in pure water. The geometrical intercept gives the data points on graph............................. 125 Figure 48. Sensor resistance as a function of pH, in alternating cycles involving alternating bubbling of CO2, purified and unpurified Ar through water........ 126 Figure 49. FTIR spectrum of the dried emeraldine base thin film.................................. 128 Figure 50. Comparison of the FTIR spectra between the dried film and the film immersed for 10 minutes in carbonic acid solution of pH = 3.95.................. 129 Figure 51. Comparison of the FTIR spectra for emeraldine base treated in HCl solution of pH = 3.88 and 1.75....................................................................... 130 Figure 52. FTIR spectrum of the EB film treated in pH = 1.11. ..................................... 131 Figure 53. Comparison of the FTIR spectra between the films equilibrated in water and in carbonic acid solution of pH = 3.95 for 10 minutes. The spectra is recorded after the acid-base treatment of the film was accomplished. .......... 132 Figure 54. Resistance of emeraldine base thin film as a function of pH of carbonic acid, generated through bubbling of various mixtures of carbon dioxide and argon in deionized water. ....................................................................... 134 Figure 55. The appearance of the second semicircle at a pH of 5.85, generated by bubbling a mixture of Ar-0.015% CO2 for 30 minutes in pure deionized water .............................................................................................. 135 Figure 56. The high frequency resistance of the emeraldine thin film as a function of carbon dioxide concentration in Ar-CO2 mixture (black) and pH response of solution to bubbling carbon dioxide (red).................................................. 136 Figure 57. Impedance of emeraldine base thin film for two fixed frequencies as a function of carbon dioxide concentration in argon-CO2 mixture; a correlation between the carbon dioxide concentration and pH generated in pure water is displayed in Figure 56. ......................................................... 137
xiv
LIST OF TABLES Table 1. Reduction elements and cathode materials employed in amperometric CO2 sensors design. ............................................................................................. 20 Table 2. Conductivity increase produced by an increase in relative humidity for emeraldine base sample equilibrated at pH = 6.0; *For the sample whose conductivity increased by several orders of magnitude the author, W.R. Salanek estimated that a response was governed by the preparation procedure which allowed for virtually removal of all the water during the dynamic pumping. The latter data represented a personal communication of W.R. Salanek to the authors of reference [192]............................................... 54 Table 3. Alternative methods of fabrication employed. 8, 12, 24 represent the polymerization time (hours); the numbers in regular type indicates air and bold type indicates N2; *-indicates that PVA with two different molecular weights (13000-23000 or 89000-98000) were used; **- indicates that PVA with three different molecular weights (13000-23000, 89000-98000 or 124-186000) were used. ....................................................................................... 65 Table 4. Measured values for the suspended sensor. ........................................................ 79 Table 5. Characteristic peaks and vibrational modes assigned to emeraldine base .......... 91 Table 6. Full width at half maxima and peak half width over peak height for emeraldine base samples heat-treated at various temperatures. .......................... 98
1
1. INTRODUCTION AND RESEARCH OBJECTIVES
1.1. Safety Issues Faced in Food Chain Industry
The current research represents a part of a project developed by Auburn
University to respond to the problems that arise in the U.S. food supply chain. The
science and engineering developed through integration of sensor and information
technologies will help accurately and rapidly identify and characterize the nature of the
food contamination. The envisioned integrated system is expected to deliver information
not only for the detection of real-time contamination (e.g. bacterial, chemical and
surface) but also for the traceability and inventory of the product (time-temperature
control).
The food industry currently relies upon end-product testing for bacterial
measurements. This consists of laboratory analysis of a sample taken from the production
line, having the shortest time necessary for the identification of microbial contamination
of six hours. The severity of microbial contamination, providing an answer to whether the
food is suitable for consumption is assessed after 48 hours of laboratory analysis. This
lengthy time interval is necessary to multiply the number of bacteria to detectable levels
of 104-108 colony forming units/ml [1,2]. During this two-day interval, millions of
pounds of infested food products have been already distributed, sold and consumed
without the problem being even detected. The most needed piece of information to the
food industry is when, where and how the problem occurs. The sooner the problem
2
source is identified, the prompter is the action taken, the less is the product waste and the
safer is the food distributed to consumers.
1.2. Carbon Dioxide Evolution Rate Measurement
Respiration or CO2 evolution represents a universal indicator for all biological
activities, rendering the determination of CO2 evolution rate a useful indicator for most of
the biological activities. The determination of CO2 evolution rate has been used over the
years for the evaluation of living microbial biomass in soils [3,4], pesticide damage on
soil microorganisms [5], decomposition of leaf litter [6], toxic effect of trace metals on
microbial populations [7,8], detection of microorganisms in blood cultures and specimens
and examination of contaminated bottled juices [9,10].
Precise and rapid determination of microbial contamination in packaged food is
often necessary for food sample analysis. The most widely employed technique for the
evaluation of microbial contamination of food is the total plate count (TPC), a lengthy
method that requires at least 48 hours to produce the results [11,12]. Alternatively, the
CO2 evolution rate, which is correlated with the real time microbial activity of
contaminated food, was determined with a commercial IR-CO2 analyzer, [13,14] and
with a microrespirometer [15]. The results of the latter methods have showed a good
correlation with the values provided by the conventional TPC method through the
cultural and sensory scores. Nevertheless, in spite of a very good correlation with
published data, both system are inadequate for miniaturization.
3
1.3. Classification of CO2 Sensors
1.3.1. Electrochemical Techniques
The electrochemical detection of carbon dioxide comprises potentiometric,
amperometric, conductometric, capacitive and coulometric techniques. A further
classification can be done according to the absorption place of carbon dioxide: in a solid
or in a liquid material. However, the amperometric, potentiometric and conductometric
electrochemical techniques involving absorption in solid ceramic materials require a high
operating temperature, in the range of 300-1100°C [16,17] and do not represent therefore
a competitor for the room temperature carbon dioxide detection sought in the present
project.
Therefore, under this section will be discussed just the liquid based carbon
dioxide sensors since they work at room temperature and represent a reason for
comparison with the emeraldine base carbon dioxide sensor developed herein.
1.3.1.1. Potentiometric Type CO2 Sensors: Description
The first attempt to build a carbon dioxide sensor based on a liquid electrolyte
was made by Stow et al. [18] in 1957, but an improved version of the sensor was
proposed by Severinghaus et al. [19] one year later. Following this discovery, other
reports of improved function sensors working on Severinghaus principle occurred over
the years [19,20-31]. In the potentiometric mode, a voltage corresponding to the desired
chemical concentration or activity (i.e. in this case pCO2) is measured with the sensitivity
of the potentiometric sensors governed by the Nernst equation (i.e. equation (1) below):
4
E = nF
RTII
CO
I
CO
p
p
2
2ln (1)
where: E is the actual reversible potential of the electrode measured in volts; n is the
charge number of the electrode reaction; F is Faraday constant (i.e. 96500C/mole); R =
8.314 J/mole×K is universal gas constant; T is the absolute temperature in Kelvin; pCO2I
and pCO2II represent the extra and intra cellular concentration of carbon dioxide
respectively. Cai and Reimers [31] showed that an effectively constant reference potential
can be guaranteed by using moderately high concentration of bicarbonate (i.e. 10-3 to 10-2
HCO3-) in the internal electrolyte reservoir.
The carbon dioxide gas diffuses through a hydrophobic gas permeable membrane
(i.e. typically ~20 µm thick) into a thin electrolyte layer (e.g. water or NaHCO3) trapped
between the membrane and the tip of a flat surface or a bulb pH electrode, until the
equilibrium is established on both sides of the membrane, as schematically displayed in
Figure 1 adapted from reference [19]. An Ag/AgCl reference electrode in contact with
sodium hydrogen carbonate solution completes the electrical circuit to the pH working
electrode. The potential difference between the two electrodes is measured by a standard
high impedance voltage amplifier and display system. The most employed material for
the gas permeable membrane is silicone rubber. This membrane material allows easy
diffusion of carbon dioxide into the internal solution, where part of it dissolves and form
CO2(aq) [26,28,30].
The set of chemical reactions involving carbon dioxide hydration within the thin
electrolyte layer are presented below, along with their respective constant of reaction.
5
Gas permeable membrane
CO2 Electrolyte thin film region
H+
H+
H2CO3 ↔ H+ + HCO3-
CO2 Sample
pH buffer
To high impedance voltage amplifier Ag/AgCl reference
electrode
Electrolyte reservoir
Glass pH electrode
H+
Glass outer body
0.5-15 mm tip
Epoxy
Figure 1. First generation Severinghaus type carbon dioxide sensor.
CO2(aq) = pCO2 × KH (2)
where KH = 30 ×10-3 M/atm is Henry’s constant [26]. The dissolved CO2 reacts with
either water or hydroxide ions, according to the reactions (3) and (4) below and forms
carbonic acid and carbonium ion respectively [26,28,30].
CO2 (aq) + H2O ↔ H2CO3 (aq) (3)
6
with KS =)(
)(
2
32
aqCO
aqCOH = 2.6 × 10-3 M/atm
CO2 (aq) + OH- ↔ HCO3- (4)
Both the carbonic acid and carbonium ion dissociate pH dependently (according to
reactions (5) and (6) below) [26,28,30].
H2CO3 ↔ HCO3- + H+ (5)
where K1 = ][
][][
32
3
COH
HCOH−+ ×
= 1.72 × 10-4 M/atm is the first dissociation constant for
carbonic acid.
HCO3- ↔ CO3
2- + H+ (6)
where K2 = ][
][][
3
2
3
−
−+ ×
HCO
COH = 5.59 × 10-11 M/atm is the second dissociation constant for
carbonic acid.
To account for the fact that CO2(aq) is in equilibrium with H2CO3(aq), the first
equilibrium is normally given by:
K1 = )]([
][][
2
3
aqCO
HCOH−+ ×
= 4.47 × 10-7 M/atm (7)
If we write the dissociation constants K1 and K2 in terms of chemical activity of each
species present, and assuming that Henry’s solubility law holds in the thin film layer, it
can be shown [19,20,31] that the carbon dioxide partial pressure pCO2 in the thin
electrolyte film is given by
pCO2 =
+
−+
+
+++
H
WHNaH
a
KK
Kaaa
21
2
21α
(8)
7
where KW = aH+ aOH
- = 10-14 M2 is the ionic product of water and α is the solubility
coefficient of carbon dioxide in the thin electrolyte layer. The above reaction (8) is valid
when a thin layer of water fills the narrow space between the tip of the pH electrode and
the membrane. However, the design employing a distilled water electrolyte produced an
unstable sensor [18], presumably because of the pH shift induced by even a slight
contamination of the distilled water electrolyte. Severinghaus et al. [19] have introduced
5-20 mM concentration of sodium hydrogen carbonate as an electrolyte in order to
stabilize the pH and increase the sensitivity. Under this circumstance, apart from the
above reactions (i.e. reactions (1) to (6)), sodium hydrogen carbonate dissociation occurs:
NaHCO3 ↔ Na+ + HCO3- (9)
Therefore, equation (9) above simplifies to
pCO2 = 1K
aaHNa
α++
(10)
Severinghaus also demonstrated that when the pCO2 in the hydrogen carbonate layer
changed to a new value, then by observing that both aNa+
and K1 are constants and cancel
out when two aH+ values are inserted and by taking logarithm of both sides of equation
(10), the relationship between the changes in pCO2 and the pH in bicarbonate solutions is
given by:
∆log pCO2 = -∆pH (11)
with ∆ standing for a change in logarithm of pCO2 and a change in pH.
In the bicarbonate layer, by definition:
∆pH = - ∆log aH+ (12)
The sensitivity of the sensor (S) is defined by Severinghaus as:
8
S = ∆log pCO2/ ∆pH (13)
Thus, if S = -1 then each order of magnitude change in pCO2 will theoretically generate a
one unit change in the pH of the bicarbonate solution which would be registered as a
change in voltage of 61.5 mV at 37°C (i.e. human body temperature), as predicted by
Nernst equation.
A problem associated with a detection set-up based on the above hydration
reactions of carbon dioxide (i.e. reactions (3) and (5)) is that the attainment of
equilibrium is very slow. In order to address this inconvenience, the carbonic anhydrase
enzyme has been added to the electrolyte solution to speed-up the reaction [26,32-34].
Following the same principle as the so-called zinc-hydroxide mechanism of CO2 catalysis
in human blood, the carbonic anhydrase catalysis in aqueous media results in the
formation of HCO3- and H+ ions starting from water and carbon dioxide. In addition,
being a very powerful catalyst, with a turnover of 106 reactions per second
[26,28,32,33,35] carbonic anhydrase facilitates a short response time of the sensor.
The typical dimensions of the first generation Severinghaus-type potentiometric
carbon dioxide sensors were 3-3.5 mm in length and 1-1.5 mm outside body diameter
[19]. The first miniaturized version of the Severinghaus-type microelectrode appeared 16
years later [21]. A schematic of a potentiometric type CO2 sensor of the most recent
generation having an internal pH electrode with a liquid ion exchanger membrane (LIX)
is presented below in Figure 2 [25,26,28,34,36]. The inner pH sensor with an ion
exchange membrane tip is positioned within the shaft of the electrode, behind the gas-
permeable membrane.
9
Ag/AgCl reference electrode
LIX membrane
Electrolyte reservoir: 2mM NaHCO3 + 0.5 M NaCl
Outer casing
Ag/AgCl wire
pH microelectrode
50-300 µm tip
pH buffer reservoir (0.7M HCl)
Silicone rubber (gas permeable) membrane
Electrolyte thin film region
Figure 2. Miniaturized version of the Severinghaus-type CO2 sensor employing liquid ion exchange membrane.
This arrangement requires an even smaller internal pH microelectrode tip, and the
subsequent placement of the latter inside the shaft of the outer casing. Beyenal et al. [28]
reported that merely 1 out of 10 microsensors built in their laboratory displays reliable
calibration curves and is useful in further utilization. Other authors [31,34,36] used a
somewhat larger sensor design, having an external tip diameter in the range of 50 to 300
µm, but these dimensions were not a result of the difficulty in achieving miniaturization.
Zhao et al. [34] in fact showed that for a certain glass material, the ratio of the surface
area of glass to the volume of solution (S/V), is inversely proportional to its tip diameter.
10
This ratio determines the extent to which the performance of a pCO2 microelectrode is
influenced by chemicals leaked from the glass wall. Thus, the smaller the tip diameter,
the larger the S/V and the greater the effect exerted by the glass material over the sensor
performance. With this respect, Zhao et al. [34] showed that if the tip of the outer casing
is less than 150 µm, then the performance of the microelectrodes depends on what type of
glass materials is used to construct the inner pH microelectrode. In their study, Zhao et al.
[34] reported that alumina silicate glass showed better results than Pyrex glass whereas
the lead glass displayed the poorest results. The above mentioned group correlated these
results to the greater weight loss of the lead glass (i.e. 3.6 mg/cm2 measured after 6 hours
exposure at 100°C to a 5% NaOH solution, compared to 1.4 mg/cm2 for Pyrex and 0.35
mg/cm2 for alumina silicate glass). When the tip of the outer electrode was further
decreased to a value below 50 µm, the performance of the sensor decreased irrespective
of the glass material used in fabrication of both the outer and inner pipette [34], the
authors suggesting that other type of glass with high chemical stability being necessary to
employ for further improvement of the system performance. Zhao et al. [34] conclusions
are supported by Cai and Reimers [31]. In the latter study, a sensor stability of two weeks
was reported for a sensor with 300 µm tip and the outer shaft made of lime glass (i.e. 1.1
mg/cm2 weight loss). However, a large tip diameter is not necessarily good in all the
aspects of detection, since it greatly improves the stability but reduces the response of the
sensor up to 14 mV/decade of µatm pCO2 for the detection at low carbon dioxide partial
pressure [34]. In an exactly opposite way, a small tip diameter improves the detection at
low CO2 partial pressure, but has a reduced lifetime to only several hours for a 4-10 µm
tip [34]. Similar results were reported by other authors [20,25,28] which suggested that
11
the external tip of the sensor needs to be smaller than 20 µm, (e.g. a value of 2 µm being
in fact reported by Hanstein et al. [26]) in order to improve the detection characteristics
(sensitivity, selectivity, response time) at low CO2 concentrations.
Initially was considered that the thickness of the gas selective membrane
represented the rate limiting factor in achieving a good response time. Instead, studies
performed by Cai et al. [31] and Beyenal et al. [28] show that the dissociation of H2CO3
is actually the rate limiting factor. Zhao et al. [34] shed more light into this problem,
showing that in the presence of carbonic anhydrase, the hydration of the CO2 gas
represents the rate-limiting-step if the membrane thickness ranges between 15 and 35 µm.
The carbon dioxide hydration equilibrium is displayed in the equation (14) below which
represents the actual equilibrium between CO2 and water, in fact the summation of
equations (3) and (5) above in the case of reasonable assumption that the equilibrium
concentration of H2CO3 and CO32- is negligible [37]):
CO2 + H2O ↔ HCO3− + H+ (14)
All the measurements performed with membranes having dimensions within this
interval resulted in obtaining identical response times for the sensor. Above 35 µm, the
response of the sensor was governed by both diffusion and rate of hydration reaction, and
it became completely dependent on the diffusion process through the membrane if the
thickness exceeded 365 µm. The group did not employ membranes with thicknesses
smaller than 15 µm, without mentioning the reason behind their choice.
The Severinghaus type potentiometric carbon dioxide sensors are relatively
unstable, displaying a reproducibility and stability of only several days. Drifts ranging
from mV/h [34] to 10 mV/h [21,25,28] were reported. In contrast with these initial
12
values, the improved Severinghaus type microelectrode for CO2 detection display a lower
drift, which is acceptable for practical applications. Published drift values in the literature
are: 0.5 mV/h by Komada et al. [36], 0.2 mV/h by Cai and Reimers [31], 0.4 mV/h by
Zhao et al. [34]. The slopes of the calibration curves reported for the latest generation
miniaturized potentiometric type carbon dioxide sensors were reported in the range of 56
to 80 mV/decade of µatm pCO2 [25-31,33,34], in fact double in value than the reported
slopes for the first miniaturized Severinghaus-type potentiometric CO2 sensor (35 to 45
mV/decade of µatm pCO2 for the latter [21-24]). Several authors [28,30,33,34] have
showed that lower values of the calibration curve slopes belong to aged electrodes and
higher values to the fresh ones.
The lowest CO2 partial pressure detected with improved miniaturized
Severinghaus-type CO2 sensors was respectively reported as 6.37 × 10-4 atm by Beyenal
et al. [28], 5.78 × 10-4 atm by Cai and Reimers [31], 4 ×10-5 atm by Hanstein et al. [26],
2 × 10-5 atm by Cammaroto et al. [33], 2 × 10-4 atm by Zhao and Cai [34], compared to
2.63 × 10-2 atm as reported by Caflisch and Carter [21] for the first generation of
miniaturized Severinghaus type CO2 microelectrode.
However, another problem that arises because of miniaturization is that the small
volume of carbonic anhydrase solution (less than 10 µl) renders the solution susceptible
to changes in compositions due to exposure to the pH microelectrode and Ag/AgCl wire.
The composition of the solution inside the pH microelectrode (the H+ selective buffer) is
extremely important since a more acidic buffer than the electrolyte buffer solution will
lead to the leak of the protons from the inner pH microelectrode solution into the
electrolyte solution containing anhydrase. This proton leak may change just slightly the
13
concentration of the anhydrase solution, but enough to create a drift since the enzyme
displayed an increase in activity with increasing pH up to pH = 10 [35]. In order to
address this issue, Hanstein et al. [26] introduced an excess quantity of sodium hydrogen
carbonate (NaHCO3) in order to fix the pH to a value around 8.3, in conformity to
Henderson-Hasselbalch equation (15) below, in fact a rearrangement of equation (5)
above. The pH value of 8.3, being different from both pK1 and pK2, warranted a sufficient
sensitivity to carbon dioxide:
pH = pK1 + log [HCO3-] / [ H2CO3] (15)
Since the first report of a potentiometric type CO2 sensor [19], the research
focused mainly on the improvement of the working electrode functions [26-
28,30,31,34,36,38-40]. So far, the most employed pH microelectrodes consisted of glass
(i.e. alumina glass [34], lead glass [34], Pyrex glass [34], lime glass [31], borosilicate
glass [26]) or Sb-SbOx materials. Despite their good sensitivity, selectivity and stability,
the shortcomings displayed by glass electrodes rest in their difficulty to miniaturize and
sometimes unsuitability for measurement inside the solid food products due mainly to
their fragile configuration. On the other hand the Sb-SbOx working electrodes suffer from
oxygen interference and a subsequent drift. Due to these disadvantages, the development
of new reliable materials to substitute for the glass (i.e. alumina silicate, Pyrex, lima,
lead) or Sb-SbOx became an urgent necessity.
Iridium oxide was proposed as a potential substitute. The sensors developed using
this novel material showed a very good dynamic range and a negligible drift [27,28,30,
39,40]. As an advantage, the sensors having the tips made of iridium oxide showed no
need for the liquid ion exchanger membrane, the iridium oxide being actually grown on
14
the tip of the iridium inner sensor as schematically showed in Figure 3 adapted from
reference [28]. In this configuration, they displayed improved stability and response time
[28,30,34,39]. A drift of 0.03 mV/h, in fact 10 times smaller than the lowest reported drift
among the sensors employing liquid ion selective membrane on their tip was reported
[28,39].
Reference electrode wire
IrO2 pH sensing tip
Iridium Oxide
Iridium wire
Glass Electrolyte: 2mM NaHCO3 + 0.5 M NaCl
Outer casing
Iridium wire
pH microelectrode
10-20 µm tip
Ag/AgCl reference electrode
Silicone rubber (gas permeable) membrane
Figure 3. Iridium oxide tip miniaturized Severinghaus type CO2 sensor.
The range of response times have been reported in seconds or minutes. Generally
the response time was placed in the range of minutes [28,31,33] but reports placing it in
the range of seconds also appeared [25,26]. Thus de Beer et al. [25] reported “between 10
15
seconds and few minutes”, while Hanstein et al. [26] reported an interval between 18 and
63 seconds, with a mean of 43 seconds. However, the response time falls in the range of 1
to 3 minutes in most of the reports. The difference in the response time can be associated
with many different factors: the composition of the internal electrolyte, the nature of the
gas selective membrane, and the dimension of the tip of the sensor. In addition, the way
to calculate the response time is an important issue that is not yet standardized. Generally,
it is calculated as the time elapsed from the beginning of the measurement to the point
when the signal reached 90% of the amplitude, but the interpretation of this interval
remains at the authors’ discretion. Moreover, the reported values represent sometimes an
average of the measurements taken with different devices.
The potentiometric methods of detection require a very strict control over
experimental variables (i.e. temperature and ionic strength) if one is concerned with
accurate measurements. Due to the logarithmic scale of the Nernst equation (1), for the
detection of a univalent ion, a one mV error generated during the reading of the electrode
potential will result in a 4% error in concentration reported. In the case of a divalent ion,
this error will be double. The potentiometric sensors, being actually pH dependent,
display interference not only from ions but also from organic and inorganic acidic or
basic volatile compounds (e.g. H2S, NO2, SO2, benzoic, cinnamic), leading to an
undesired potential change. In addition, the presence of microorganisms that could
produce a pH shift represented another problem to consider in designing the
potentiometric type sensors for carbon dioxide. Last but not least important is the
inability of Severinghaus-type potentiometric sensors to perform extended time
monitoring of carbon dioxide concentration. This fact is due to the bulk of electrolyte
16
which would theoretically play no part in the equilibration process and is envisioned to
serve just as a reservoir and electrical conductor. All active processes take place only in
the thin electrolyte layer at the tip of the pH microelectrode, and is therefore important to
keep the sensor immersed in a stable reference gas or liquid solution between readings to
allow in this way the sensor to work always from a stable pH reference point in the
sodium bicarbonate layer and to return to the same point after the CO2 measurement. If
the sensor is maintained too long in a sample solution, the pCO2 in the bulk of the sodium
bicarbonate electrolyte reservoir will begin to come into equilibrium with the pCO2 of the
sample and this event will destroy the stability of the sensor.
1.3.1.2. Amperometric Type CO2 Sensors
Another method used in the detection of carbon dioxide is the amperometric
mode [33,41-71]. In the amperometric mode, the gas of interest is allowed to pass
through a gas-permeable membrane and diffuse in the inner electrolyte solution until
equilibrium is established on the both sides of the membrane. The formation of a metal
salt (e.g. chloride, sulfate, phosphate, depending on the type of electrolyte employed)
occurs due to the pH change induced by the carbon dioxide dissolution in the electrolyte.
The metal salt can be reduced to the electrode through a quasi-reversible path, the current
raised during the reduction process being proportional to CO2 concentration in the
electrolyte solution and therefore easily correlated to the CO2 partial pressure in the
analyte medium.
Evans et al. [46,47] showed that although the response of their sensor resulted
from a pH change of the electrolyte, the chemistry involved, the nature of the response,
17
and the design of the sensor differ from the Severinghaus type carbon dioxide electrode.
It was initially thought that since the current density is dependent upon the rate of
reduction of Cu2+ ions at cathode, then the increase in the number of ions through
increasing the concentration of copper diamine complex would increase the sensitivity of
the device [46,47]. What was discovered in reality was actually much more complicated.
A low concentration of the copper diamine complex (i.e. 2-3 mM) rendered a higher
sensitivity towards carbon dioxide (up to maximum 10% CO2), whereas a too high
concentration (i.e. 50 mM) resulted in a limited current density due to the formation of a
precipitate in the solution (identified as copper hydroxide). In the same time, a too low
complex concentration (i.e. 0.5 mM) gave a non-linear current density vs. % CO2 curve,
with a short linear portion suitable for the detection of CO2 concentrations less than 1%.
However, the sensor showed poor stability mainly due to the strict requirement placed
over the selection of copper diamine complex concentration. In designing a system to
work on the same principle as presented by Evans et al. [46,47], Fasching et al. [50]
chose non-volatile buffer substances N,N-bis(2-hydroxyethyl)-2-aminoethansulfoneacide,
N-tris (hydroxymethyl) methyl-2-aminoethansulfoneacide, and N,N-bis (2-
hydroxyethyl)glycine and showed that these ones form more stable ligand compounds
with copper ions. The choice of the ligand is in fact more complex than just its volatile
behavior. This is because the reaction of copper with hydroxyl ions results in the
formation of copper hydroxide, according to the equation (16) below:
Cu2+ + 2OH- → Cu (OH)2 (16)
The poor solubility of the copper hydroxide (1.6 × 10-19 mol/l) corroborated with
the ease of crystallization can lead to a depletion of the free copper in the electrolyte and
18
degradation of the sensor reproducibility and stability. The formation of copper
hydroxide is favored by high free ion concentrations at pHs greater than 7, therefore in
order to keep the Cu2+ concentration in solution low, Fasching et al. [50] proposed the
use of ligands with almost complete complex formation at higher pHs. Dissociation of
copper complexes that takes place in the pH range between 7.7 and 6 resulted in the
release of unbounded copper ions, corresponding to a pCO2 in the range of 10-3 atm up to
one atm. A schematic of an amperometric sensor of the latest generation developed by
Fasching et al. [50] is presented below in Figure 4 adapted from reference [50].
CO2
CO2
analyte electrolyte
cathode anode
Gas permeable membrane
e-
CO2 + H2O ↔ H+ + HCO3- ↔2H+ + CO3
2- CuL2
2+ + 4H+ ↔Cu2+ + 2H2L2+
Cu2+ + 2Cl- + e- → CuCl2-
CuCl2-→ Cu2+ + 2Cl- + e-
Figure 4. Sensing concept of the amperometric CO2 sensor.
1.3.1.2.1. Reducing Specie; Electrode Material
The first report about the reduction of carbon dioxide at platinum electrode dates
more than 40 years [55,56]. Carbon dioxide represented in fact the species that was in
19
most of the cases directly reduced [41,43,45,49,52-57,62]. Unfortunately, CO2 is not
electroactive over a larger potential range, and CO2 reduction wave overlaps with H+
reduction in aqueous solutions [41,46-48]. Due to this inconvenience, aprotic solvents
(e.g. DMSO, N,N-dimethylformamide, acetonitrile) were investigated [41,43,45,51-
53,57]. Wadhawan et al. [70] showed that compared to water, DMSO offers the
advantage of a significantly greater solubility of CO2, corroborated with the ability to
withstand wider cathodic potential until the solvent breakdown occurs. Detailed
information about the cathode material and reduction element as well as the respective
reference is presented in the Table 1.
20
Table 1. Reduction elements and cathode materials employed in amperometric CO2
sensors design.
Reduction Element Cathode Materials Reference
Carbon Dioxide Mercury 53
Carbon Dioxide Gold 52,53
Carbon Dioxide Gold microdisc (φ = 10 µm ) 52
Carbon Dioxide Amalgamated gold 52
Carbon Dioxide Amalgamated platinum electrode 52,55,56
Carbon Dioxide Platinum 57
Carbon Dioxide Thin film platinum 45
Carbon Dioxide Platinum microdisc sealed in glass 43
Cu2+ Miniaturized platinum cathode 50
Cu2+ PTFE-sprayed-platinum porous electrode 46,47
Hydrogen Platinum Oxide 41
Carbon Dioxide Rhodium 62
Carbon Dioxide Miniaturized rhodium (thickness = 60 nm) 49
Carbon Dioxide Silver 41
Hydrogen Ruthenium Oxide 48
1.3.1.2.2. Gas Permeable Membrane
The gas permeable membrane that allows easy diffusion of carbon dioxide from
the analyte medium into the inner electrolyte solution is in most of the cases fabricated
from silicone rubber, but Teflon, gold-coated Teflon and platinum oxide sputtered PTFE
have also been employed [41,43,45,49]. As an alternative to these hydrophobic
membranes, Cammaroto et al. [33] employed an immobilized enzyme (carbonic
anhydrase) coupled with a redox mediator (p-benzoquinone) which used CO2 as a
21
substrate, replacing in this way the electrolyte with a suspension of carbonic anhydrase in
Tris-HCl buffer of pH = 8.3.
Qian et al. [43] found that oxygen presence in the solvent generated a prewave on
the voltammogram of CO2 reduction while causing a successively increasing of the peak-
shaped current, which could rise up to ten times as much as the normal limiting current.
This increase suggested complicated changes in the overall reaction. The source of
oxygen could be either the oxygen gas dissolved in the liquid electrolyte or that contained
in the gas sample to be detected. Therefore, since the oxygen interference could not be
corrected by simple mathematics the purpose of the amperometric sensor designers was
to avoid the presence of oxygen in the electrolyte by shielding the working electrode
from O2 interference. Simultaneous electrochemical detection of carbon dioxide and
oxygen in nonaqueous solutions like DMSO or acetonitrile was challenged by the easier
reduction displayed by the oxygen compared to carbon dioxide in those solvents
[41,45,46]. Due to the easy reduction of oxygen in aprotic solvents, it is therefore
impossible to detect the concentration of a small percentage of CO2 in the presence of a
much larger O2 generated current. The oxygen interference is caused by the proximity of
the reduction potentials of the two gases, and the overlapping of the reduction waves in
aprotic solvents occurs at far less negative potentials as compared to aqueous electrolytes
[42,45,46]. At those negative potentials necessary for CO2 reduction, the reduction
product of O2 is a superoxide ion that interferes with CO2 measurement and accumulates
inside the solution [41,42]. In aqueous solvents however, no oxygen interference is
expected if the potential of the cathode is controlled to an appropriate value (0.15 and
0.22 V vs. Ag/AgCl for potassium chloride and potassium bromide electrolyte solutions
22
respectively) [47]. Until now, three attempts were made to circumvent this superoxide
interference effect by designing gas-phase O2-CO2 sensors using macro-cathodes.
In first of these attempts, Albery et al. [42] employed double membrane-double
solvent layer sensor. The outer gold metallized membrane where O2 was reduced in an
aqueous electrolyte layer buffered at pH = 5 was polarized at -0.7 V. The current raised
after O2 reduction at gold metallized membrane was proportional to O2 concentration, the
O2 filtering offered by the membrane being estimated at 99%. The carbon dioxide passed
through the outer membrane, through the first layer of electrolyte and through another
Teflon membrane, reaching an inner compartment composed of DMSO and silver macro-
cathode as electrolyte/electrode system respectively. The CO2 was reduced to CO2.- at
silver electrode. The construction itself was however complicated and led to a slow
response time of the sensor due to a double membrane barrier. In addition, Hahn [20]
communicated in his review paper that no other group could reproduce Albery et al. [42]
results.
In a second attempt, the electrochemical filter technique for O2 was not followed
any longer and a pulse titration technique employed [68-71] instead. Unfortunately, the
set-up proposed by Albery et al. [68,69] was suitable only for gas phase detection and the
gold macrocathode (i.e. 1.5 mm in diameter) using a pulse titration technique, was not
suitable for miniaturization and involved a complex set of chemical reaction in the
electrolyte layer [68,69]. The set of chemical reactions proposed by Roberts et al. [72] is
in accordance with Albery et al. [68,69] as well as other groups finding [70,71]:
1) The reductive pulse that lasts typically 200 ms [20,69] performed at a potential
where CO2 reduction is inactive:
23
O2 + e-→ O2.- (17)
2) The titration step in which some of the O2- reacts with any CO2 present:
2 O2- + 2 CO2- →C2O62- + O2 (18)
3) The recovery step (oxidation of the untitrated O2- back to O2), accomplished by
switching the electrode to an oxidizing potential (typically lasts 350 ms);
O2- → O2 + e- (19)
The difference between the generated and the recovered O2 is a measure of the
concentration of CO2 in electrolyte. Unfortunately, the reduction pathway exemplified in
the reactions (17)-(19) above has the shortcoming of regenerating O2, which is further
reduced during the first voltage pulse, leading to an increase of the observed overall O2
reduction signal. This event makes necessary a complex set of mathematical model to
deconvolute the true O2 and CO2 concentrations [69].
The third attempt to solve the superoxide interference problem was performed by
Qian et al. [43]. The group employed a porous hydrophobic platinum catalyzed Teflon
membrane electrode to remove oxygen in the first step and performed a measurement of
CO2 at a platinum microdisc in a following step, the system designed consisting therefore
in a two separate amperometric sensors combined. Qian et al. [43] showed that the first
membrane electrode completely consumes the oxygen, allowing only the carbon dioxide
to reach the second amperometric sensor consisting of DMSO and platinum micro disk of
60 µm diameter. As with the other two methods attempted, this third one was also suited
only for gaseous samples analysis. Due to the different transport times of the two gases,
the overall system designed by Qian et al. [43] displayed a 90% change in O2 and CO2 in
15 and 35 seconds respectively and could not be employed in breath-by-breath analysis
24
that required a sensor with a response time less than 0.1 second. As an improvement to
the setup proposed by Qian et al. [43], Hahn et al. [65-67] reported a sensor capable to
measure simultaneously O2 and CO2 at unshielded gold microdisc cathodes in DMSO
electrolyte, creating in this way a single amperometric sensor cell. The current generated
after the carbon dioxide reduction was obtained by subtracting the O2 background current
from the total current. The authors showed that the reduction of each specific gas was
independent of the amount of the other, and the reduction took place at different
potentials making the system display minimal interference.
1.3.1.2.3. Electron Transfer Studies
The attention of electrochemists focused towards understanding the nature of the
electron transfer process between reducing specie and cathode material. Sawyer’s group
[52,53] employed chronopotentiometry to extract information about the reaction
occurring at electrode in amperometric carbon dioxide sensors. In chronopotentiometry, a
constant current is passed through an electrolytic cell and the potential of the working
electrode is monitored. The variation of this potential with time takes the form of a wave,
and information about the reactions occurring at the electrode can be obtained from the
shape and duration of this wave. This analysis led the group conclude that the gold
electrode allowed just one electron reduction whereas the platinum electrode gave two
electrons. In a later study, Eggins et al. [57] confirmed the nature of one electron
reduction for gold cathode, suggested by Sawyer’s group [52,53]. Haynes et al. [53]
suggested that the reduction at mercury electrode is a diffusion controlled one electron
overall reduction process, dependent also on the applied potential. Thus one electron per
25
mole of CO2 is consumed at less negative potentials (-2.3 V vs. SCE) and two electrons
per mole of CO2 are consumed at more negative potentials (-2.5 V vs. SCE). The two-
electron value is believed to result from catalytic solvent reduction by an intermediate
species at the more negative applied potential. All these studies (i.e. [52,53,57]) were
consistent in concluding that whether or not water was present in the aprotic system
employed, the reduction of carbon dioxide involved the production of CO2.- in the first
step, followed thereafter by a multitude of competing possibilities. These various
pathways that followed the CO2.- generation depend strongly on the experimental factors
like current density and diffusion layer concentration and thickness, leading to
uncertainty in extrapolating the macro scale electrolysis results to the micro scale and
vice-versa.
1.3.1.2.4. Reduction of O2 and CO2 at Micro-electrodes
The most important difficulties encountered in designing an amperometric sensor
for carbon dioxide detection were represented by the instability of the electrolyte and the
overlapping of the reduction potentials of oxygen and carbon dioxide. In addition, the
electrode dimension represented a factor of influence. With this respect, Hahn [20]
showed that the CO2 reduction waves completely disappear in the presence of oxygen at
large size electrodes. Moreover, Zhou et al. [45] reported that microelectrodes are
preferable over the regular size electrodes. The microelectrodes display a much smaller
ohmic drop over the electrolyte, making the measurements possible even using highly
resistive non-aqueous media [20,40,58,70]. The microelectrodes also display a reduced
capacitive component of the total current, which leads to an increase of the signal-to-
26
noise ratio in transient faradaic measurements. Finally, the steady states of mass transfer
are rapidly established, which makes the microelectrodes convenient for performing
steady state current measurements. On the other hand, the response profile of the CO2
microsensors depends on the thickness of the diffusion-limiting electrolyte thin layer on
the microelectrode surface. The use of microelectrodes was later shown to be necessary
for the simultaneous detection of CO2 and O2 since it was believed that the kinetics of the
intermediate reactions that preceded the formation of the peroxy dicarbonate anion was
outrun by the much faster diffusional loss at the electrode-electrolyte interface [20,65-
67]. By employing miniaturized electrodes the current signal that arises at cathode is
independent of the presence of another gas. Regarding the material choice for the
miniaturized electrodes, Hahn et al. [65-67] described gold as a better reductive element
than the investigated platinum or glassy carbon disc for the same 10 µm diameter. On the
other hand Zhou et al. [71] proposed a platinum cathode microdisc of 60 µm diameter.
1.3.1.2.5. Characteristic Values Reported
Zhou et al. [45] reported the lower detection limit as 0.4% CO2 but other groups
recorded even lower values: 0.1% by Fasching et al. [50], 0.04-0.1% and 0.025% by
Ishiji et al. [41,48]. The response time of the carbon dioxide amperometric sensor is
much lower compared to that of potentiometric counterparts. Response times ranging
from 10-50 seconds [33,45,46] to 2 minutes [50] were reported, but also a very low value
of only 5 seconds reported by Cammoroto et al. [33] in a configuration including
carbonic anhydrase as a catalyst of the CO2 hydration reaction. The sensor stability was
reported in the range of hours [45], days [50] or even weeks [41].
27
1.3.1.2.6. Conclusions
Because the electrode reactions in amperometric mode are not fully reversible, but
at most quasi reversible [46,47,50], the overall electrochemical CO2 process consumes
CO2 from the sample. Consequently, any liquid or gas sample displaying the same CO2
partial pressure will register different output when analyzed with the same amperometric
sensor. There is therefore a so-called “liquid-gas difference” for the amperometric mode
of detection of carbon dioxide and this is a major disadvantage of the amperometric mode
over the rival potentiometric Severinghaus-type technique. However, this issue was
somewhat resolved in the configuration employing micro-electrodes for carbon dioxide
reduction, since they consume a negligible small quantity of analyte in which they are
measuring and also produce a minimal quantity of bi-products.
1.3.1.3. Conductometric Type CO2 Sensors
Conductometry represents another electrochemical technique employed for
carbon dioxide sensing in which the conductance of the solution rather than the electric
potential or current is measured [37, 73-80]. The conductance is a function of the pH of
the solution and therefore dependent on CO2 partial pressure.
The very first attempt to make use of the conductance of an electrolyte and to
correlate this result with carbon dioxide concentration belonged to Cain et al. [73], who
sought to extract information about the content of carbon in steel. The method employed
by the group consisted of passage of the carbon dioxide gas resulted after the direct
combustion of a sample of steel into a Ba(OH)2 solution of known electrical resistance.
28
The absorption of carbon dioxide gas in the solution resulted in precipitation of barium
according to the equation (20) and a subsequent increase in electrical resistance that
could be correlated with the amount of gas dissolved and therefore to the initial content
of carbon in the sample metal.
Ba(OH)2 + CO2 = BaCO3 + H2O (20)
A similar approach was used by Holm-Jensen [74] to determine the CO2 concentration in
air by employing a solution of barium hydroxide and measure its change in conductivity
as a result of a passage of a stream of air. Other authors [75,76] introduced the
measurement of the conductance of the deionized water after the absorption of carbon
dioxide in order to determine the amount of gas.
Lis et al. [77] reported a CO2 sensor working on the principle of electrical
conductivity by performing the measurement of the gas dissolved into a potassium
hydroxide solution. Nevertheless, the sensor was rather big, with the dimensions in
centimeters and required a very long preparation time prior to use (i.e. 2 weeks of
treatment in distilled water) to clean the impurities on the glassy sensing head.
In one of the first attempts to measure carbon dioxide dissolved in liquid sample,
Himpler et al. [37] employed a thin layer of water separated from the analyte sample by a
dialysis membrane. The set-up represented actually a modified Severinghaus type sensor
in which the conductance of solution was measured instead of a pH response as a result
of carbon dioxide hydrolysis in aqueous electrolyte. By doing this, the problem of analyte
depletion was eliminated, since the Severinghaus-type sensor does not necessitate the
existence of an infinite sample reservoir of carbon dioxide. The one minute response time
of the sensor depended on the kinetics of the hydrolysis of CO2 in the thin water layer.
29
This response time is comparable to the response time of the first generation miniaturized
Severinghaus-type potentiometric carbon dioxide sensor [21-24] and the sensitivity
covered the analytical range of interest for blood analysis. However, Himpler et al. [37]
detected a slow increase in conductance with time of exposure at the same carbon dioxide
partial pressure. Proving that this small drift was not caused by the inadequate sample
mixing or slow responding electronics, the group suggested that the problem might have
been generated by the diffusion within the unstirred thin water layer inside the
membrane. Another disadvantage of the development was that the time required for the
sensor to detect a low to high CO2 concentration transition was shorter than the reversed
way.
Bruckenstein and Symanski [78] solved the drift problem by introducing a thick
ion-exchanger layer bed separated from the thin water layer through a thin screen. The
square of the conductance of the solution was then proportional to the carbon dioxide
partial pressure in the water layer. The sensor responded to a range of concentrations of
CO2 from 0.45 to 5%. Unfortunately, the design showed unacceptable sensitivity to
movement due probably to segregation or shifting of the particles composing the ion-
exchanger bed. As a disadvantage, it was shown that the lighter anion-exchanger resin
beads (responsible for HCO3- removal), which composed the ion-exchanger bed settled
more slowly time than the heavier cation-exchanger beads. As a result, the conductance
slowly increased in time due to a longer time required for carbon dioxide to diffuse down
inside the ion-exchanger bed in order to be consumed [78].
One of the first reports regarding the use of polymers conductivity for CO2
detection belongs by Saikai et al. [81] who employed a mixture of solid polyethylene
30
glycol and alkali carbonate. The sensor developed had a CO2 dependent conductivity
with a trade off in the lower detection limit, which could not be lowered below 0.5%
CO2.
A disadvantage of the conductometric method is the lack of selectivity of the
system if more than one species is involved in the detection mechanism. Attempts to
miniaturize the conductometric sensors showed that the stability is greatly affected by the
dissolution of the thin electrolyte water layer with the easily volatile species presented as
impurities on the sensor walls [37,79,80].
1.3.1.4. Coulometric Type CO2 Sensors
Coulometric titration [23,24,82-92] to measure the pH of the solution represents
another electrochemical technique employed for carbon dioxide detection. The design of
ISFET follows the same principle as MOSFET (metal oxide field effect transistor), with
only difference, that the metal gate of the MOSFET is replaced by a metal of a reference
electrode, while the liquid in which this electrode is present creates the contact with the
gate insulator. Both devices have characteristic the same electrical circuit, displayed in
Figure 5.
Vgs G +
-Id
D
S
Vds
Figure 5. Schematic of the electrical diagram for ISFET.
A coulometric sensor is in fact an integrated sensor-actuator device that measures
the pH changes induced to an electrolyte solution (e.g. water) by ions generation (i.e.
31
either H+ or OH- depending on the electrochemical reaction) at a generating actuator (e.g.
a coulometric gold electrode) with the aid of a sensor (i.e. a pH sensing ion sensitive field
effect transistor, ISFET). A schematic of a coulometric sensor is displayed in Figure 6
adapted from reference [84].
p-Si
n+(source) n+ (drain)
Inner electrolyte layer
Membrane (gate)
Au
Ta2O5
SiO2
Id
Figure 6. Schematic of the coulometric carbon dioxide sensor.
The coulometry allows one to develop a fixed relation between the coulombs
applied to the actuator and the amount of generated ions, provided that the following
conditions are fulfilled:
1) The stoichiometry of the electrode reaction is known;
2) No side reaction occur in the system;
3) The current efficiency of the electrode reactions approaches 100%.
The ions generated at the actuator can be considered as the titrant, with which a
coulometric titration can be accomplished. In this way, the pH-sensitive ISFET is used as
the indicator electrode for the detection of the end-point of the titration curve. The main
advantage of the system is that it is suitable for performing acid-base titrations in
nanoliter ranges in time of seconds. Another appealing feature of coulometry is that the
32
output of the sensor is provided by its own dimensions and not by the sensitivity of the
working electrode, therefore a one time calibration is expected to be sufficient for an
envisioned long time stability of the sensor.
A problem associated with the naturally occurring titration of carbon dioxide is
the slow nature of it. Both reactions, the one involving carbonic acid formation and the
following one involving carbon dioxide titration with hydroxyl ions are slow-a complete
equilibrium occurring in about 5 minutes. However, a back titration if employed can
speed up the reactions, mainly because the speed of the titration reaction of carbon
dioxide with hydroxyl ions (i.e. reaction (4) above) is pH dependent. To do this, the
excess of hydroxyl ions is introduced into electrolyte (i.e. carbonate) solution, raising the
pH to about 10.5 to 11 pH units, values at which the reaction of CO2 with OH- is
accomplished in few seconds, consuming all the carbon dioxide. The excess OH- is now
titrated in a regular way and the protons generated gave information about the amount of
carbon dioxide originally present in solution.
In one of the first studies involving the latter method, Van der Schoot and
Bergveld [84] presented an ISFET based potentiometric CO2 sensor having a relatively
good stability over a period of seven weeks. The drift reported (i.e. 0.05 mV/day) is much
lower than the drift characteristic to usual potentiometric sensors.
The membrane thickness has a great influence over the response time of a
coulometric sensor [84,91]. The presence of the membrane is necessary because the gate
insulators materials normally used in ISFET preparation undergo surface ion exchange
process especially due to H+ presence that makes them prone to drift especially in the
presence of alkali metal ions [91]. In order to solve this issue, membranes are chosen
33
from materials able to incorporate ionophores. By doing this, the interface of membrane
to solution displays a potential selectivity for some ionic specie of interest, while
preserving the permeability toward neutral specie (i.e. weak acids and bases, water
vapors, CO2, O2, organic molecules).
Unfortunately, problems appear in this configuration due to the hydrophilicity of
the membrane that accumulates a fragmented water layer on the surface, permitting some
acid base reactions between neutral specie to take place at the interface and alter the
potential. An alternative remedy to this problem has been proposed by Van den Vlekkert
et al. [92] who introduced a thin water buffer layer between the gate insulator and the
polymeric membrane. However, this set-up was not amenable to miniaturization because
of the restricted buffer capacity of the layer and the subsequent leak due to high osmotic
pressure. A viable solution nevertheless is the introduction of a thin Ag/AgCl or AgCl
layer to the membrane surface to eliminate the neutral specie pH mediated sensitivity
[93,94].
Coulometric sensors, as well as the other type of liquid based carbon dioxide
sensors briefly described above (i.e. potentiometric, amperometric, conductometric),
despite the very good achievements with respect to the response time and sensitivity, lack
long term reliability mainly due to the existence of a liquid solution. The incorporation of
a liquid, not only complicates the cell design and severely restricts the miniaturization but
also affects the durability and integration of the sensor with microelectronic interrogation
devices.
34
1.3.2. Spectroscopic Techniques
Most of the sensor designs based on spectroscopic techniques employ optical
fibers [95-115]. This development is possible since carbon dioxide displays a strong
infrared absorption band extending from 4.2 to 4.4 µm. Optical CO2 sensors are divided
into two categories. One is based on the optical detection of a color change in a pH
indicator dye, mostly employed being thymolsulfonphthalein (thymol blue) and
phenolsulfonphthalein (phenol red) [96,102-105,107,109,111]. The other configuration
works on the principle of the CO2 induced fluorescence change of a luminescent dye such
as 1-hydroxypyrene trisulfonate [95-97,102,114].
The spectroscopic techniques involving dye-solvent solutions and optical fibers
are advantageous if one is concerned with simultaneous determination of oxygen and
carbon dioxide. The absorption bands in the visible region of the two elements do not
overlap, providing no cross-interference between the two gases [101]. In addition to this,
the major advantage of the optical fiber based detection rests in the ability to detect
virtually the full range of carbon dioxide partial pressures, from almost zero to one atm.
Optical fibers mode of detection was employed also in conjunction with
microdialysis technique [102], the system developed being able to concomitantly monitor
hydrogen, oxygen and carbon dioxide partial pressures in blood and buffer solutions
[102]. The system employed closed loops of micro-flow streams of reagents in microliter
per minute, a simple optical system and a method of achieving steady microliter per
minute flow rates, proved to have an excellent resolution in the range of interest for
oxygen and carbon dioxide for medical applications. The above approach offered also a
35
response time of only 4.8 min, 1.8 min, and 2.1 min for pH, pCO2, and pO2 measurement
respectively [102].
A response time of only 40 seconds and a dynamic range covering all carbon
dioxide concentrations (from virtually 0% to 100%) was reported for a configuration
consisting of a carbonic anhydrase introduced into the immobilized phosphate buffer
layer placed at the distal end of an optical fiber [96]. In this configuration the measuring
principle rests on a CO2-modulated proton transfer from the inner buffer to the photo
excited tris[2-(2-pyrazinyl) thiazole] ruthenium(II) cation, immobilized onto an anionic
dextran gel.
A solid state carbon dioxide sensor based on colorimetric detection principle
comprises a dye immobilized in a water-immiscible polymer film or plastic material
(polymer + plasticizer) [96,103-113,115]. The role of plasticizer is to ensure the diffusion
of carbon dioxide in and out of the solid compound. The difficulty in achieving such a
configuration is represented by the incorporation of water into the solid material
employed (i.e. plastic or polymer film), to serve as a medium where the carbon dioxide
hydration reactions take place. The nonpolar, hydrophobic nature of both the polymer and
plastic medium impedes the incorporation of water. In addition, the solid medium should
be designed in such a way to display little or no leaching of any of the sensing
components. The above inconveniences can be circumvented to some degree by using a
phase-transfer agent. Mills et al. [107] employing a sensor with a phase transfer agent,
obtained a very low response time (less than 3 seconds) but at the same time they noticed
that prolonged use of the sensing films lead to a deterioration of the sensing element.
36
Rocchia et al. [112] used a different approach for their carbon dioxide sensor
working in diffuse-reflectance FTIR mode. Instead of using a plasticizer and a reference
dye, the group introduced an amine-modified silica gel that interacted with carbon
dioxide and created carbamates species, which were successively detected by infrared
spectroscopy. However the setup was not developed as a portable device and in addition
the repeatability of the sensing mechanisms was questionable, since the silica gel needed
reheating at temperatures greater than 40ºC to liberate the bound carbon dioxide.
A very low response time of only one second was reported for a configuration
employing the color change mechanism of a covalently immobilized dye in silicone
rubbers [104]. This type of sensor is capable of measuring carbon dioxide both in gases
(to which they respond with a t90 of 0.1 to 1 s) and in aqueous sample solutions (with t90
ranging from 2 to 8 min). The authors attribute the longer response time in water sample
to the presence of water-filled micro droplets in the membrane, which constitute a
reservoir for the gas-equilibrated sample. However, the sensor design needed careful
storage and a preparation procedure (i.e. water equilibration) before use in aqueous
sample.
Unfortunately, the spectroscopic techniques are not well suited for the
determination of CO2 partial pressure in liquids due to interference from other absorbing
species, such as water vapor and carbon monoxide [101,104]. In addition, the
achievement of equilibrium of carbon dioxide hydration and dehydration reactions is
slow, leading to the necessity of inserting the carbonic anhydrase enzyme as a reaction
catalyst. However, as already mentioned in the electrochemical techniques section,
carbonic anhydrase solution becomes unstable due to the presence of microorganisms,
37
leading to a slow and undesirable pH shift. For the systems employing dye and solvent,
some authors [101,104] observed the existence of a small signal drift caused by a small
evaporation of the solvent and tried to circumvent it by simple replacement of the dye-
solvent solution with a fresh one at regular time intervals. Moreover, for detection of
carbon dioxide in liquid media, these types of sensors were reported to have a lower
response time when measured virtually low to high CO2 variation compared to the
reversed way [96,110].
1.3.3. Other Methods of Detection
Quarts crystal microbalance (QCM) [116-123] and surface acoustic wave (SAW)
devices [124-128] have been also employed for carbon dioxide detection.
The QCM method works on the principle of a quartz crystal oscillating frequency
monitoring, subsequent to carbon dioxide interaction with the crystal coating. The
technique takes advantage of the reactivity of carbon dioxide with amines, imines or
zeolites at room temperature. However, the last two classes of compounds showed
unsatisfactory results since zeolites displayed adherence problem with the crystal surface,
whereas imines proved unsuitable due to degradation in time [116-123]. Among the
amines, the N,N,N’,N’-tetrakis(-2-hydroxyethyl) (THEED) showed good response [120-
123]. However, this coating displayed low stability in time (i.e. the oscillating frequency
changed in time, and consequently the sensitivity towards CO2 fluctuated) and in addition
presented interference from other volatile species (i.e. SO2, NH3). The
tetramethylammonium fluoride tetrahydrate (TMAF) coating on the other hand displayed
a better stability, but in the same time a lower sensitivity than THEED [120-123].
38
Surface acoustic wave sensors employ polymer coatings capable to absorb carbon
dioxide and change the oscillating frequency ∆f of the SAW device according to the
equation (21)
∆f = (k1 + k2) f02 m/A (21)
where: f0 is the operating frequency, k1 and k2 are material constants and m/A is the film
mass per unit area.
The SAW devices were reported to work at room temperature [127-128], although
an operating temperature in the range of 70 to 100 ºC offered higher sensitivity [124-
126]. However, the room temperature detection is severely interfered by humidity and
difficulty in finding an appropriate polymeric material to allow a high partition
coefficient between the gas phase and the sensor surface. In fact Grate et al. [129]
showed that the working principle is more complex than just a simple mass change of the
coating. The above-mentioned group identified the swelling-induced modulus change
effect as a major contribution to SAW vapor response. Therefore, the SAW sensor
responses to gas vapors of interest are in reality multiplied beyond the effect of vapor
mass alone. Grate et al. [129] compared the partition coefficients of the same vapor
detected by SAW and by gas-liquid chromatography (GLC) and found always that
partition coefficient in SAW case, KSAW was greater than KGLC, sometimes by 4-6 times.
Ruling out all the other possible influencing factors (i.e. volume, density, temperature
change, etc.), the group concluded that the only factor that could explain such a
difference is the dynamic modulus variation. Therefore, in selecting a material for a SAW
device, the choice of material with high dynamic modulus is preferable in order to obtain
the greatest potential for modulus effects.
39
Nevertheless, the main difficulty in devising polymeric materials for CO2
detection is represented by the lack of discrimination displayed by these materials with
respect to other interfering gases and water vapor. Thus, silicon based polymers such as
poly (γ-amino propyl ethoxy / propyl ethoxy silane) (PAPP) and poly (γ-amino propyl
ethoxy / octadecyl ethoxy silane) (PAPO) have been employed in order to reduce the
interference from water vapor [117]. Other coatings containing silicon compounds (e.g.
amino propyl trimethoxy silane (AMO) and propyl trimethoxy silane (PTMS)) have been
used in different CO2 types of sensors based on capacitance response [130]. In addition,
mixtures of polymers have been employed to improve the sensor performance [128].
1.4. Conductive Polymers as Carbon Dioxide Sensors
1.4.1. Conductive Polymers: Description
Conducting polymers are polymers that can be modified in such a way that their
electrical conductivity varies over several orders of magnitude. Alan MacDiarmid, Hideki
Shirakawa and Alan Heeger discovered the conductive polymers almost 30 years ago and
shown that they can be doped over the full range from insulator to the metal state [131].
The classical saturated polymers, in which all of the four valence electrons of
carbon are used in covalent bonds, appear in the form of insulating materials and
therefore are uninteresting from the point of view of electronic materials. Conjugated
polymers however, have an electronic configuration that is fundamentally different. In
the latter case, the chemical bonding leads to one unpaired electron (i.e. the π-electron)
per atom bonding. This property becomes further interesting because the π-bonding in
which the carbon orbitals are in sp2pz configuration and in which the orbitals of adjacent
40
carbon atoms along the chemical backbone overlap, leads to electron delocalization along
the polymer chain. The conductivity achieved through the doping or charge injection is
due to the introduction of charge carriers into the π-electron system that leads to the π-
electron delocalization and the movement of electric charge along the polymer chains.
1.4.2. Insulator to Metal Transition Process
The first two conductive polymers discovered were sulfur nitride, (-SN-)n [132-
134] and polyacetylene, (-CH-)n [131], but the latter one turn out to be the one
extensively investigated. In polyacetylene, each carbon atom is linked through a σ-bond
to each of the two neighboring carbon atoms and the single hydrogen atom in the
repeated unit cell. Moreover, the hydrogen atoms overlap laterally through a π-bond,
leading thus to a structure with one π-electron available for conduction for each carbon
atom of the polymer backbone. This ideal structure of polyacetylene with equal length of
the carbon-carbon bond would imply the existence of a metallic state. In reality, due to
Peierls instability, the polyacetylene structure is dimerized [131,135-143], leading to a
structure that has characteristic the existence of two carbon atoms per repeat unit (-
CH=CH-)n.
A consequence of this arrangement is the division of the initial π-band into two
bands: π and π∗. Since only two electrons (spin up and spin down) are allowed to fill each
band, the consequence is the attainment of fully occupancy of the π-band and an empty
π∗-band. The energy gap between the highest occupied level in the π-band and the lowest
unoccupied level in the π∗-band represents in this way the energy gap (Eg) that has to be
41
overcome in order to make the polymer conductive [131,135-142]. The existence of the
energy gap between a lower fully occupied band and an upper empty band represent a
characteristic feature of the insulators or semiconductors materials. This energy gap of
potentially conductive polymers depends upon the molecular structure of the repeat unit
and therefore gives the chemists the possibility to perform its control by design at the
molecular level. In addition, the discovery of non-linear excitations in this class of
conductive polymers, (i.e. solitons and confined soliton pairs (polarons and bi-polarons)
allowed the interconnection between electronic and chemical structure [131,135-158].
There are in fact several methods through which conductive polymers can
increase their conductivity:
1) Electrochemical doping;
2) Chemical doping;
3) Photo-excitation;
4) Charge injection at the metal-insulator-semiconductor interface.
However, the reversible doping in conductive polymers that allows a full control of
conductivity at the entire range from insulators to the metal state can be achieved by
chemical and electrochemical doping [131,135-142,146,159-172]. The conductivity
achieved through photo-excitation and charge injection at the metal-semiconductor
interface is just transient and lasts as long as the electrical or optical doping stimuli
remain active. Wegner et al. [140] proposed a general equation of the doping process in
conductive polymers:
Pn ↔ Pn·+ A- ↔ Pn
2+ 2A- (22)
42
for the oxidation reaction, the reverse reaction standing for the reduction process. Pn is
defined as a periodic segment of a polymer chain that exchanges a full charge with either
a redox partner (i.e. in the chemical doping) or with the electrode of an electrochemical
cell (i.e. in the electrochemical doping) and is thereby transferred to its hypothetical first
or second redox state [140,173,174]. This charge transfer process leads to the formation
of a cation (anion for the reversed reaction) in the first redox reaction and dication
(dianion) for the second electron transfer.
The ion radical is called “soliton” in the case of movement free and independent
of each other of both the charge and spin along a single chain without being subjected to
relaxation or trapping. This ion radical is instead denominated as “polaron” when it is
linked to an elastic bond distortion and consequently cannot move without exchange of
energy. The ion radical that appears because of the second redox reaction is called
“bipolaron” and it is subjected to disproportionation according to reaction (23) below:
2 Pn·+ ↔ Pn
0 + Pn2+ (23)
depending on the temperature and the type of binding to the appropriate counterion.
During the doping process, the electrochemical potential (Fermi level) is moved
either by a redox reaction or by an acid-base reaction into an energy region where there is
a high density of electronic states. In the same time charge neutrality is maintained by the
introduction of the counter-ions in the vicinity of the chain, rendering the conductive
polymers salt-type characteristics [131,134-144,145,146,156-159,161-164,166,167,169,
171,173]. After the doping process, the electrical conductivity results from the existence
of charge carriers (introduced through the doping process) and from the ability of those
carriers to travel along the “highway” of overlapped π-orbitals.
43
In conclusion, the doped conjugated conducting polymers have good conductivity
characteristics primary due to the following reasons:
1) The doping process results in an introduction of carriers into the electronic
structure. Moreover, since every repeat unit represents a potential redox site, the
conductive polymers can be doped either p-type (oxidized) or n-type (reduced) to a high
charge carriers density [131,135-142,144-146,152,155,156,159,161-164,166,167,169,
171,173].
2) Charge carrier delocalization and its respective mobility along the polymer
backbone are developed due to the attraction of an electron in one repeat unit to the
nuclei of the neighboring units. This process of delocalization and the subsequent two-
dimensional mobility of charge carriers along the polymer’s chemical backbone are
extended also in three dimensions through inter-chain electron transfer [131,135-142,
144-146,152,155-159,161-164,166,167,169,171,173].
This group of conductive polymers has been employed for the detection of CO2
[81,175-181], and represent a solution for the detection of carbon dioxide at room
temperature.
1.4.3. Polyaniline: Description
Among the conductive polymers, one of the most investigated is polyaniline [145,
157-159,161,162,164,167,182-184]. MacDiarmid et al. [160-162] proposed that the
reduced form of a repeat unit in polyaniline contains two benzenoid rings, (e.g.
N H
N H
), whereas the oxidized form of the repeat unit contains one
44
benzenoid ring and one quinoid ring (e.g.
N N
), with none of
these two in the form presented above carrying any electric charge. The transformation of
a fully reduced into a fully oxidized repeat unit involves oxidation (i.e. loss of a hydrogen
atom, in fact an electron and a proton), as schematically showed in Figure 7-adapted from
reference [160].
oxidation
reduction
N H
N H
N N +2H+ + 2e-
Reduced form Reduced form Oxidized form
Figure 7. Oxidation of a fully reduced aniline repeat unit.
Chiang et al. [160] summarized the set of possible repeat units that could be
found in polyaniline:
1) Unprotonated repeat units:
a) Completely reduced unprotonated polyaniline base repeat unit (1A)
N H
N H
b) Completely oxidized unprotonated polyaniline base repeat unit (2A)
N N
2) Protonated repeat units:
a) Completely reduced monoprotonated polyaniline salt repeat unit (1S’)
A-
H+
N H
N H
45
b) Completely reduced diprotonated polyaniline salt repeat unit (1S’’)
A-
H+ A-
H+
N H
N H
c) Completely oxidized monoprotonated polyaniline salt repeat unit (2S’)
A-
H+
N N
d) Completely oxidized diprotonated polyaniline salt repeat unit (2S’’)
A-
H+ A-
H+
N N
Polyaniline forms a wide range of oxidation states, from fully reduced form
(leucoemeraldine) up to the fully oxidized form (pernigraniline). The leucoemeraldine
chains of polyaniline consists of aromatic benzenoid rings, which are bonded to amine
nitrogens, with each of the amines bonded to only one hydrogen atom. On the other hand,
the pernigraniline form is composed of alternating benzenoid and quinoid rings, which
are bonded only to imine nitrogens, with no hydrogen atoms bonding to the chain. The
half-oxidized form of polyaniline is referred as emeraldine, and consists of an equal
number of oxidized and reduced units (1:1 ratio of amine: imine nitrogens) and 3:1 ratio
of benzenoid: quinoid rings.
A schematic of three-oxidation forms of polyaniline is presented in Figure 8.
46
2x N N N N PB
2x N N EB N
H
N
H
2x LB N
H
N
H
N
H
N
H
Figure 8. Polyaniline in various oxidation states.
In reality, polyaniline exists in more than three forms of oxidations. About 100
years ago, Green and Woodhead [185] were able to synthesize, isolate and analyze five
distinct forms of polyaniline base, showing that the minimum number of (C6H4)N groups
necessary to permit an interconversion between those forms is eight. The above five
specie were synthesized by Green and Woodhead [185] by means of chemical oxidation
or reduction with agents having characteristic standard reduction potentials. Green and
Woodhead [185] showed that in fact this polymer could exist in a continuum of oxidation
states ranging all the way from fully reduced to a fully oxidized material
Depending on the pH of the protonating solution and the prior history of the
sample (i.e. the degree of oxidation induced by the fabrication method), every polyaniline
can undergo protonation to some extent. Huang et al. [162] showed that the extent of
protonation decreases with increasing the extent of oxidation in the chemical backbone,
thus pernigraniline shows little or essentially no protonation in acidic solution of pH
greater than 1. Therefore, theoretically speaking, all the possible forms in which the
polyaniline exists can be converted to their corresponding mono or diprotonated salt by
treatment with an acid of appropriate strength [160-162]. However, studies performed on
47
the properties of the resulting salts showed that only emeraldine base could achieve a
metallic-like conductivity when reaching its protonated salt form [160-162,167,172]. The
chemical environment renders the emeraldine conductive through the protonation of
imine nitrogen and/or deprotonation of amine nitrogen sites. Besides, protonation studies
performed over the polyaniline base powders [157,158,160-169,171,186] showed that in
spite of the fact that the amine nitrogen atoms present a greater base strength than the
imine nitrogen counterparts, almost all of the imine nitrogen atoms protonate first. This
preferential protonation of imine nitrogen atoms in emeraldine base is probably due to the
better resonance stabilized structure that results in polymer chain.
MacDiarmid et al. [164,168,182] showed that an emeraldine sample prepared in
air has 60% of the sites in oxidized form, whereas a sample prepared under inert
atmosphere displays just 40% oxidized sites. It appears therefore that an increase in the
number of oxidized sites would conduct to an increase in protonation and a higher
conductivity. However, this is not the case, and as stated above, a sample oxidized in
excess of 50% displays a lower conductivity than a sample 50% oxidized [164]. The
same argument is valid for a sample substantially more reduced (e.g. prepared in air, with
40% oxidized sites). An explanation of this lower conductivity of emeraldine salts
displaying a ratio of reduced to oxidized units different that one would rest in the
generation of internal redox processes that occur in order to maximize the extent of
conjugation in the delocalized polysemiquinone radical cations [182]. The latter radical
cations would therefore localize and form metallic islands, whereas the excess oxidized
and reduced groups would thus migrate and localize in the insulating beaches [182].
48
For a fully protonated, resonance-stabilized emeraldine salt chain, MacDiarmid et
al. [160-162] have showed that all nitrogen atoms, all carbon-nitrogen bonds and all the
C6H4 rings would be identical. Thus, all the nitrogen atoms in the ring would bear a + 0.5
electric charge, all the nitrogen atoms would be intermediate between an amine nitrogen
and an imine nitrogen, all the C-N bonds would be intermediate between a single and a
double bond and finally all the C6H4 rings would be intermediate between benzenoid and
quinoid rings. Figure 9-adapted from reference [162]-depicts the resonance-stabilized
form of emeraldine salt as proposed by Huang et al. [162]. In such a stabilized form, the
highly conjugated π–electron system creates the premises for recording the metallic-like
conductivities (up to 400 S/cm) for the emeraldine salt [183,184,187-189].
In addition, Paul et al. [186] and MacDiarmid et al. [160-162] showed that the
maximum conductivity of the emeraldine salt polymer occurs when all the imine nitrogen
units in the chain are in the (2S’’) form above exemplified. On the other hand, if the
protonated form of emeraldine salt occurs in the monoprotonated amine nitrogen site
(1S’’), then the conjugation will be severely interrupted, and the resonance stabilization
of the backbone affected [160-162,166,186,190].
N H
N H
N
H
N H
N
H
A-
N H
H A-
N
A-
A-
+0.5
+0.5
+0.5
+0.5
+0.5
+0.5
+0.5
Figure 9. The resonance-stabilized form of the emeraldine salt polymer.
49
The sequence of structural changes in the emeraldine form of polyaniline after
doping with a protonating acid is depicted in Figure 10-adapted from reference [137]-
with all the steps briefly mentioned in the insert. During the doping process, the
alternating polymer-like chemical structure of insulating emeraldine base is transformed
by the proton-induced spin unpairing mechanism into a chemical structure with only one
unpaired spin per repeat unit but with no change in the number of electrons [146,147,149,
150,160-162]. The outcome of this process is the formation of a polaron lattice that has a
smaller repeat unit compared to the original emeraldine base and which allows the
delocalization of the π-electrons along the chemical backbone and thus metallic-like
conductivity.
50
Reaction with protonic acid ads 2H+
n N H+
N H+
N H
N H
N H
n
⋅⋅⋅⋅⋅⋅⋅⋅ ⋅⋅⋅⋅⋅⋅⋅⋅ ⋅⋅⋅⋅⋅⋅⋅⋅ ⋅⋅⋅⋅⋅⋅⋅⋅ ⋅⋅⋅⋅⋅⋅⋅⋅ N N
H N H
N N H
n H+
N N H
N H
N N H H+
Geometrical (bond length) relaxation (quinoid to benzenoid)
Redistribution of charge and spin; halving of unit cell
N H+
N H n
Figure 10. Protonation induced charge unpairing in polyaniline. The counterion display is omitted from the schematic.
Studies over the properties displayed by the emeraldine resulted after different
methods of preparation revealed that there are two distinct classes [222-224]:
1) Class I is characteristic to emeraldine generated from solution in the protonated
salt form (ES-І), and then converted to the respective unprotonated base form (EB-I)
through treatment with a strong base. This class of emeraldine polymers displays
substantial crystallinity in the ES-I form but is essentially amorphous in the EB-I form.
51
2) Class II materials are obtained when the polymer is synthesized in the
unprotonated base form (EB-II) and subsequently transformed to the protonated salt (ES-
II) through treatment with a strong acid. The polymer displays crystalline behavior in the
salt form, with a different crystal structure than ES-I, and a partially crystalline structure
in the non-protonated, EB-II form. In addition, the EB-II crystal structure can be
generated by dissolving the EB-I type in an appropriate solvent (e.g. tetrahydrofuran
(THF), dymethyl sulfoxyde (DMSO) or N-methyl-2-pyrrolidone (NMP) and casting a
film from the respective solution.
Polyaniline was proposed by Ogura et al. [175] as a material able to undergo
variation in conductivity in response to a change in both the concentration of CO2 and
humidity of the surrounding atmosphere. Water vapor, in fact, affects the doping and
chemical structure of conducting polymers [191-195], which results in a change in
conductivity, therefore allowing for the use of conductive polymers as humidity sensors
[196-199]. A schematic of self-doping of polyaniline occurring in the presence of water is
presented Figure 11-adapted from [175], following a mechanism proposed by Ogura et
al. [175,177] in the case of a polyaniline derivative.
⋅ + ⋅ +
N H
N H
N H
N H
N H
High H2O:
COO- COO-
Low H2O: N N H
N H
N N H
COOH COOH
Figure 11. Self doping of emeraldine base.
52
The influence of water on polyaniline has been investigated in studies performed
on polyaniline derivatives like poly(o-phenilene diamine) PoPD, poly(o-aminophenol)
PoAP, poly(m-phenilene diamine) (PmPD) and poly(o-toluidine) PoTD. In all the above
cases the variation of conductivity of polyaniline derivative polymers as a function of
water vapor is not linear. The linearity can be achieved in some particular cases if the
polyaniline derivatives are added to poly(vinyl alcohol), generating thereafter a
composite polymer mixture [ 175,177-180,195,198-203]. Ogura et al. [198] suggested
that a linear response of the polymeric mixtures appeared when the polyaniline
derivatives displayed fibrillar morphology, whereas non-linear response was produced by
a granular morphology. The growth of fibril-like structure was possible due to a higher
solubility of some polyaniline derivatives in their respective solvent (e.g. DMSO). In
addition, the fibril-like morphology allowed for a greater dispersion of water within the
polymer composite [198,199].
An interesting observation was that the polyaniline derivatives provided the
sensitivity to water vapor concentration although they were minor constituents in the
polymer composite [175,177-180,198-200,204,205]. For example, additions of just 0.23-
0.28 vol.% polyaniline to PVA provided a linear response and the response became non-
linear when more than 4.3 vol.% polyaniline was added [199]. In addition to this
property, the conductivity of the composite polymer changed by 1 to 2 orders of
magnitude when the relative humidity increased from 11% to 98% [175,177,196,198,
199].
Another study performed by Ogura et al. [199] on different polyaniline
derivatives revealed that composite polymers in which the water molecules were strongly
53
bound to the PVA were less affected by the variations of external humidity. Ogura
suggested that this property aroused because the conductivity of the polymer was
dependent on interaction between a protonic acid (e.g. toluene-p-sulfonic acid) and
polyaniline. During this interaction, the weakly bound water molecules promoted the
protonation of imine-nitrogen sites and produced a subsequent increase in conductivity. A
decrease in humidity reduced the amount of weakly bound water molecules in the
composite matrix chains, an event that was followed by a subsequent decrease in
protonation and electrical conductivity. In the extreme cases when the external humidity
level was very low, all the water became strongly bound to the PVA and polyaniline
formed the insulating emeraldine base. Ogura et al. [199] showed that the conductivity
loss was less prominent for the polyaniline derivatives that displayed strong attraction to
water molecules, rendering these composites a lesser dependence of their conductivity to
the relative humidity. With this respect, Ogura et al. [175,177-180] showed that for a
desirable output of the sensor, the water molecules should be as much bound as possible
to the PVA matrix.
Travers et al. [191-193] and Angelopoulos et al. [194] on the other hand,
attempted to explain the effect of water on electrical conductivity of polyaniline. The
above-mentioned groups also observed an increase in conductivity of both “dried” and
“exposed to water” emeraldine base that was suddenly exposed to higher level of water
vapor pressure. The increase in conductivity was found to be dependent upon the pH
level at which the emeraldine base was previously equilibrated. In Table 2 are presented
the published values, for emeraldine base samples equilibrated at pH = 6.0.
54
Table 2. Conductivity increase produced by an increase in relative humidity for emeraldine base sample equilibrated at pH = 6.0; *For the sample whose conductivity
increased by several orders of magnitude the author, W.R. Salanek estimated that a response was governed by the preparation procedure which allowed for virtually removal
of all the water during the dynamic pumping. The latter data represented a personal communication of W.R. Salanek to the authors of reference [192].
Conductivity Change RH Change Reference
Several orders of magnitude N/A* Reference 6 in 192
1-2 orders of magnitude 11% to 98% 175,177,196,198,199
1.5 orders of magnitude 0 to 100% 191
62% 0 to 65% 192
17% 0 to 17% 194
22% 0 to 17% 192
25% 21% to 42% 192
Studies performed over the magnetic, optical and transport properties of
polyaniline [145,161,163,165,172,174,190,206] described the material as composed of
segregated metallic islands (i.e. ~250 Å diameter) embedded in an insulating matrix.
Angelopoulos et al. [194] noticed a rapid increase in conductivity upon exposure to water
vapor followed by a slow decrease in conductivity under dynamic vacuum, and ascribed
the event to the above mentioned morphology. Angelopoulos has proposed that when
emeraldine base is exposed to humidity, the absorbed water decreases the insulating
interparticle resistance before diffusing slowly within the metallic islands, the latter step
being responsible later for only a minor increase in conductivity. During the dynamic
pumping, however, the water in the interparticle region is replenished as nearly as fast as
it is removed, leading to a slow increase in resistivity during the drying step.
However, the model proposed by Angelopoulos et al. [194] would imply a
necessity of different kinetics for the absorption and desorption process of water
55
molecules in polyaniline, otherwise the time response of the electrical conductivity of the
material would be identical in both stages (i.e. exposure to water and dynamic pumping).
This shortcoming of the theory proposed by Angelopoulos leads to the idea that the
bound and free water molecules play a role in this process. Travers et al. [191-193]
attempted to explain the mechanism through which the presence of water molecules
favors conduction in polyaniline, but do not explain why the conductivity change is fast
in the adsorption and slow in the desorption process.
Following the model and notation proposed by MacDiarmid, Travers et al. [191-
193] ascribed the reduced forms of polyaniline (i.e. the 1A and 1S sites) to NH and NH2+
sites respectively, and the oxidized forms (i.e. the 2A and 2S sites) to N= and NH+= sites.
The cross-over between forms in the same group of oxidation state (i.e. between the
reduced forms and oxidized forms respectively) follows acid base reactions with
characteristic constants pK1 and pK2 respectively, as it is schematically presented in
Figure 12- adapted from [191,192].
56
pH
NH+=
N=
NH2+
NH
pK1=
6.5
pK2 =
2.5
reduction
oxidation a)
b)
0
1
2
3
4
5
6
7
8
9
10
A
B
NH2+
NH
N=
N+=
Figure 12. a) Schematic representation of the acid-base and redox reactions between different forms of polyaniline; b) Thermodynamical stability domains of the pH for the
different forms of polyaniline.
The reduced form 1 is then a mixture of sites NH and NH2+, and the oxidized form
2 is a mixture of sites N= and NH+=. The relative weight proportion between sites in the
same form is governed by the pH at which the sample was previously equilibrated.
Regarding the weight ratio of the two forms (i.e. 1 to 2), this is dependent upon the
sample preparation procedure (i.e. amount of oxidant introduced, the place of preparation
and time of mixing of the polymerization ingredients). All the latter conditions influence
57
the relative ratio of the reduced to oxidized sites. In summary, an emeraldine sample
displays all the above sites enumerated by Travers and Nechtschein [191,192] in a
proportion dependent upon both the redox state and the pH of the medium in which the
sample was equilibrated prior to the conductivity measurements.
The process of conduction in polyaniline consists of an electron transfer, i.e. an
electron removal from either site of the reduced form 1 (i.e. NH and NH2+) and its
subsequent transfer to a site of the oxidized form, 2 (i.e. N= and NH+=). Figure 12b
displays the two possible routes of the electron transfer. If the electron follows the route
A, then the process consists of a pure electron transfer:
-at the origin site:
NH – e- → NH+= (24)
-at the destination site:
NH+= + e- → NH (25)
Unfortunately, the origin and destination sites in the route A do not lie under the same pH
domain and therefore, the thermodynamical instability makes the electron transfer
process followed by route A highly improbable. Route B displays both sites under the
same domain for a pH level below 2.5, therefore corresponding to the same
thermodynamical stability. However, in this case, the protonation level of the two sites is
different and the site NH2+ cannot lose an electron without losing a proton. The necessary
proton in the electron transfer reaction of route B is borrowed from the water molecules
near the NH2+ sites:
NH2+ + H2O → NH + H3O
+ (26)
58
Then the electron is removed from the NH site as described above, such that the water
molecules help just unpin the charge carried by the NH2+ sites by temporarily accepting a
proton. The conduction based mechanism proposed by Travers and Nechtschein
[191,192] rests therefore in electron hopping between localized states, with a proton
exchange-assisted conduction of electrons, (denominated by the group as “PEACE”).
However, even though it seems to explain reasonably well the conduction
mechanism in emeraldine base under the influence of humidity, the theory proposed by
Travers et al. [191,193] is challenged by the impedance spectroscopy study performed by
Javadi et al. [207]. In the latter study, the conductivity of emeraldine base was
investigated as a function of frequency and exposure to humidity in the range between
D.C. and 10-1010 Hz. The conductivity was found independent of frequency, even though
this range of frequency spans from below to above the frequency responsible for proton
exchange within the polymer. This fact led Javadi et al. [207] to conclude that the
PEACE mechanism is not suitable for explaining the conductivity-humidity dependence
of emeraldine base. Comparing the mechanisms proposed by Angelopoulos and Javadi
[194, 207] on the one hand and Travers and Nechtschein [191,192] on the other hand,
MacDiarmid [182] considers that the more likely origin of the water influence over the
electrical conductivity in emeraldine rests in the effect of water on the beaches between
the metallic islands, although the mechanism of this process is not yet well understood.
1.5. Impedance Spectroscopy: Description
In the present project, the conductivity of the composite conductive polymer film
was investigated using impedance spectroscopy, a technique employed also by Zuo et al.
59
[208] in the case of polyaniline. Impedance spectroscopy implies the application of a
small perturbation in the form of either a potential or current. The perturbation is a single
sine wave or a superposition of a number of sine waves with different frequencies. From
the applied perturbation and the subsequent measured response, the magnitude of the
impedance and phase shift is determined. Since the technique is called spectroscopy the
parameters are measured as a function of frequency of the applied perturbation. The
voltage excitation in the present case is of the form: U (ω) = U0 × e i ω t and the recorded
A.C. current response of the form I (ω) = I0 × ei ω t+δ. The impedance Z(ω) is therefore
defined as a sum of a real and an imaginary part according to the equation (27) below:
( ) ( )( ) "'sincos iZZiZZeZwdI
wdUwZ i +=+=== θθθ (27)
An impedance Z(w) = Z′ + iZ″ is a vector quantity and can be plotted in the plane with
either polar or rectangular coordinates, as displayed in Figure 13.
X axis 0
Z″
Z
θ
Z′ Re (Z)
Y axis Im (Z)
Figure 13. Impedance plotted as a planar vector using rectangular and polar coordinates.
The system is further described based on series and/or parallel circuits in terms of
resistance, capacitance, constant phase element, Warburg element, inductance, etc.
60
Impedance spectroscopy helps understanding the processes that occur at
interfaces, especially the changes in physical properties of the system (i.e.
crystallographic, mechanical, compositional, electrical) as well as changes in electrical
properties (i.e. polarization) by their effects on the electrical conductivity of the system.
Macdonald [209] showed that the distribution of interfaces is a particular feature of solid-
state electrolytic cells, where the junction between the electrode and electrolyte is
considerably more complex than in the case of aqueous cells. In addition, in the solid
electrolyte case, each interface will polarize in its unique way when the system is
subjected to an applied potential difference. The rate at which a polarized region changes
when the applied voltage is reversed is characteristic feature of each type of interface:
slow for chemical reactions at the triple phase contacts between atmosphere, electrode
and electrolyte, and relatively faster across the grain boundaries in the polycrystalline
electrodes [209].
Impedance spectroscopy offers the advantage that the conductivity measurement
is taken over a range of frequencies, which can allow for obtaining information on
different conduction species and paths. These pieces of information are important in
understanding and controlling the conduction mechanisms. Moreover, the results may
indicate parameters, which should be used in a sensor interrogation for an optimum
response. As an example, the results may show that higher sensitivity or better selectivity
occurs if the measurement is made at a high frequency, fact that would require the
presence of an oscillating circuit along with the sensor.
The parameters derived from an impedance spectroscopy spectrum fall generally
into two categories:
61
1) Those pertinent only to material itself such as conductivity, dielectric constant,
mobility of charges, equilibrium concentration of charged species and bulk-generation
recombination rates;
2) Those pertinent to an electrode-material interface, such as absorption-reaction rate
constants, capacitance of the interface region and diffusion coefficient of neutral species
in the electrode.
Capitalizing on the advantages displayed by the impedance spectroscopy
technique, however, can be difficult due to possible ambiguities in the results
interpretation. This fact arises from the choice of the equivalent circuit, which consists of
a sum of ideally lumped-constant properties. In reality, all electrolytic cells are
distributed in space, and their microscopic properties may be in addition independently
distributed, rendering the equivalent circuit ineffective in interpreting the electrical
response. It is often found that Z(ω) cannot be well approximated by the impedance of an
equivalent circuit containing only a finite number of ordinary lumped-constant elements
(resistances, capacitances and inductances). Therefore, in order to take care of the
distributed properties, the distributed impedance elements (so called constant phase
elements-CPE) are introduced in the equivalent circuit. The resistance R in the equivalent
circuit stands for the lump resistance of the conductive path, therefore a given resistor in
the circuit may stand for either the bulk conductivity in the material or the chemical step
associated with an electrode reaction. A capacitance and inductance can be attributed to
regions of space charge polarization as well as with specific adsorption and
electrocrystallization process at electrode. The CPE element can account for the
distribution in space of the microscopic material properties, for distribution of the
62
conductive paths independently to one another and for the nonuniformity of the contact
between electrodes and material. Macdonald [209] showed that an equivalent circuit
involving only a finite number of ordinary lumped-constant elements could not
adequately approximate the electrical impedance. Each particular CPE element has
associated with it both resistor and capacitor values due to the resistance distribution over
a specific material volume.
Impedance spectroscopy is the technique used in the current project to determine
the electrical conductivity of the film and to assess the way this conductivity changes
with time. The technique gives information on the way the electrical conductivity
depends on CO2 and H2O partial pressures.
63
1.6. Project Objective
The objective of the current project is to evaluate the detection of carbon dioxide
by measuring the conductivity change of the composite thin film of emeraldine base and
poly(vinyl alcohol). In a first attempt, the results reported by Ogura et al. [175,177-180,
198,199] will be investigated, employing a sensor developed in a similar configuration.
The assessment of the sensor performance will be accomplished by studying the effect
displayed by CO2 partial pressure on the transport properties, sensitivity, response time,
selectivity and repeatability of polyaniline based thin film. In addition, the effect of the
humidity level on the sensitivity and selectivity of the sensor will be evaluated. If the
results of these studies will prove unsatisfactory, other routes to improve the sensor
performance will be attempted with the aim of improving the CO2 detection
characteristics and in this way elucidating the possible applications of this sensor.
64
2. EXPERIMENTAL SECTION
2.1. Suspended Sensor
2.1.1. Conductive Polymer Preparation
The preparation method of the polymer composite follows closely the procedure
reported by Ogura et al. [177]. Polyaniline was synthesized by mixing 100 mL of 0.15 M
aniline solution and 100 mL of 0.1 M p-toluene sulfonic acid (TSA) solution containing
0.15 M (NH4)2S2O8, and continuously stirring the mixture for 8 hours. The resulting
precipitate was separated by filtration through filter paper No. 1 (Watmann Int.), then
rinsed several times, first with methanol and second with a solution of 0.1 M TSA. The
precipitate was dried at room temperature under vacuum for a period of 24 hours to
obtain a dark green powder, in the form of TSA-doped polyaniline. The dried cake
collected from the filter paper was treated in 3% NH4OH aqueous solution for 8 hours in
order to be converted to the insulating emeraldine-base polyaniline form. The base
solution was filtered through filter paper No.1 (Watmann Int.) and the resulting dark blue
precipitate dried for 24 hours under vacuum at room temperature. The dried emeraldine
base powder was collected from the filter paper and stored in a closed vial in a desiccator.
A sample of emeraldine base was extracted from the stock polymer and a heat-treatment
process performed. The heat-treatment process was accomplished in a vacuum oven
(VWR Scientific Products, Model 1410M) by inserting the sample in the oven preheated
at 365°C and keeping it isothermally for 1 hour in helium gas atmosphere.
65
Stock solutions of emeraldine base-polyaniline and poly (vinyl alcohol) were
prepared by dissolving 1 mg of EB-PAni and 6 mg of PVA in 10 mL of N-methyl-
pyrrolidone (NMP), respectively. Samples were extracted from the stock solutions and a
polymer composite mixture in ratio 1 to 5, EB-PAni to PVA generated.
Alternative methods of fabrication were also investigated. Variations in the
procedures include the polymerization time (8, 12 and 24 hours), the cover gas during the
fabrication (in a fume hood in air or in glove box under nitrogen), the heat treatment of
the emeraldine base powders (no heat-treatment or heat-treated for one hour at 365ºC in
helium atmosphere), the EB-NMP solution concentration (0.01%, 0.1% and 1%) and the
EB-NMP to PVA-NMP ratio (1/5, 1/4, 1/2). The specific conditions used are summarized
in Table 3. The sensor performance was essentially the same for all the fabrication
conditions investigated.
Table 3. Alternative methods of fabrication employed. 8, 12, 24 represent the polymerization time (hours); the numbers in regular type indicates air and bold type
indicates N2; *-indicates that PVA with two different molecular weights (13000-23000 or 89000-98000) were used; **- indicates that PVA with three different molecular weights
(13000-23000, 89000-98000 or 124-186000) were used. 0.06% PVA-NMP 1% PVA-NMP
EB-NMP/PVA-NMP ratio EB-NMP/PVA-NMP ratio Heat
Treatment
EB-NMP concentration
(%) 1/5 1/4 1/2 1/5 1/4 1/2 0.01 8*,12,24* 8,24 8,12,24 8 8 8 0.1 8 - 8 - - - No HT 1 8**,24** 8, 24 8,24 8** - 8**,24**
0.01 8,12,24* 8 8,24 8 - 8 0.1 - - 8 8 - - 365°C 1 8,12,24* 8, 24 8,24 8** 8 8,24
66
2.1.2. Polymer Film Deposition and Interrogation
The electrode used for conductivity measurements was fabricated from titanium-
gold sputtered in a comb-shape configuration on quartz substrate. The pattern consisted
of a total number of 50 fingers, 25 on each side. A schematic of the sensor is presented in
Figure 14 below. Each finger was 4 mm in length (f) and 200 nm in height (h) and
presented an interdigitated space (d) of 40 µm to the adjacent fingers. Compared to the
interdigitated electrode utilized in the current project, Ogura’s group [177] comb-shaped
microelectrode presented an interdigitated spacing of only 5 micrometers, which is about
one order of magnitude smaller than that used in this work.
The interdigitated electrode was rigorously cleaned with acetone and distilled
water and then introduced for a period of 30 min in “Piranha solution” (i.e. 3:1 ratio of
50% H2SO4 and 30% H2O2 aqueous solutions respectively), following a cleaning method
reported by Revzin et al. [210].
Polymer
Film
h
d
f
Electrode A
Electrode B
TOP
VIEW
SIDE
VIEW
Substrate
Figure 14. The sketch of the interdigitated electrode.
A quantity of 1 µL of stock solution was deposited on the interdigitated electrode
in a dip-coating approach that was also followed by Ogura et al. [175,177-180,198,199].
67
The electrode having the polymer composite deposited was then placed inside a closed
glass chamber and slowly dried under flowing argon gas atmosphere.
2.1.3. Experimental Setup
A schematic of the experimental arrangement is presented in Figure 15 . The glass
chamber was partially filled with saturated solution of a salt that fixed the humidity level
to a steady value. The humidity level fixed by supersaturated salt solution was displayed
by the humidity sensor around the following means: ~60% for NaCl, ~30% for MgCl2,
~4.5% for LiCl and ~1.5% for LiBr.
Humidity
Sensor Interdigitated
Electrode
Gas In
Electrical wires
Gas Out
LiBr; RH = ~1.5%
LiCl; RH = ~4.5%
MgCl2; RH = ~30%
NaCl; RH = ~60%
Supersaturated
Salt Solution
Glass Cell
Figure 15. The setup utilized in conductivity measurements for suspended sensor.
68
The interdigitated electrode covered with the cast polymer composite film was
connected with two alligator clips and suspended inside the glass chamber about two
millimeters above the level of the supersaturated salt solution. All the time the relative
humidity was monitored inside the glass chamber with a HMP35E humidity probe of
Vaisala HM 138 humidity sensor (0.01 accuracy).
The other ends of wires were connected to the terminal of the Impedance Gain
Phase Analyzer (Solartron model SI 1260) and the glass chamber was sealed. The first
gas flowed inside the glass chamber was pure argon (99.9%, Air Gas Co.) and the
impedance of the deposited polymer film measured by the impedance gain phase analyzer
using a frequency range between 3.2 × 107 to 1 Hz. The measured data was fitted with the
equivalent circuit displayed in Figure 16 and the conductivity of the film calculated
inserting the total the resistance (i.e. R = R1 + R2) given by the equivalent circuit into
equation (28):
Rfh
d
×××=
49σ (28)
where f, h and d represent the finger length, height and interdigitated spacing
respectively, given in Figure 14, and 49 represents the number of conductive paths
between 50 fingers. The contribution to conductivity given by the material above the
finger height and between the finger ends and the side of interdigitated electrode was
neglected.
69
RC C1
R1
C2
R2
Figure 16. The equivalent circuit schematic associated with the measured data in the suspended sensor case; This circuit will also be employed in the immersed sensor case.
In Figure 16, R1 and R2 represents respectively the two variable resistances of the
equivalent circuit that fits the measured data, whereas C1 and C2 represent the
capacitances. The contact resistance RC was kept constant at 4.5 ohm during the fittings.
Following the film resistance stabilization to a constant value, the flow of pure
argon gas was suspended and a mixture of Ar + 5% CO2 from a separate tank (Air Gas
Co.) was flowed inside the glass chamber. The Ar + 5% CO2 gas mixture was flowed
until the resistance of the thin film stabilized to a new value, and then a new cycle was
restarted by flowing pure argon gas.
2.1.4. Experimental Conditions for Material Analysis
Fourier transformed infrared spectroscopy (Perkin Elmer Spectrum GX FTIR
Spectrometer) was performed on powdery samples of emeraldine base polyaniline in KBr
mode. Equal portions of emeraldine base from the stock powder were distributed in
individual vials and a heat treatment performed for each sample. The heat treatment
consisted in maintaining each sample in oven for one hour in helium atmosphere at
temperatures of 100°C, 150°C, 200°C, 250°C, 300°C or 365°C.
UV-Visible spectrophotometry analysis (CARY 3E UV-Visible
Spectrophotometer) was performed by adding two drops of 3% emeraldine base polymer
solution in NMP in a 10 mL cuvete of pure NMP.
70
Differential scanning calorimetry analysis (TA Instruments 2910 Differential
Scanning Calorimeter) was performed on emeraldine base powder samples. The analyzed
sample was kept isothermal for 30 minutes at 100°C and then ramped at 20°C/min to
400°C. After cooling down to the room temperature, a second run of the experiment was
restarted in the same described conditions.
X-ray diffraction (Rigaku DMaxB Diffractometer) was performed on emeraldine
base sample powders heat-treated for one hour in a helium atmosphere at temperatures of
100ºC, 150ºC, 175ºC, 205ºC, 225ºC, 245ºC, 275ºC, 300ºC or 365ºC. The scan range
(2θ) was 5-45° at a scan rate of 5° min-1. To calculate the full width half maxima
(FWHM) value, a horizontal base line was drawn at the base of the peak in each XRD
spectrum and the horizontal width at half distance from the peak base to the peak crest
measured. In calculating the half width over height value (HW/H), the ratio of half the
width of the base line over the full height of the peak was computed. The peak height was
considered as the length spanning from the base line to the peak crest.
For the assessment of the amount of water incorporated by poly(vinyl alcohol),
three glass slides (Corning, 22 × 30 mm) were scrupulously cleaned with acetone, warm
water and soap, rinsed with deionized water and then introduced for 30 minutes in a Petri
dish containing Piranha solution. The Piranha solution was in the same concentration and
ratio as the one used for interdigitated electrodes cleanup procedure [210]. The cleanup
procedure was finalized through abundant deionized water rinse. The cleaned slides were
placed on a clean Petri dish and dried at a temperature of 100°C for one hour on the top
of a hot plate inside a fume hood. The glass slides were then removed from the heat
source and a quantity of 1.5 mL of a solution of 0.06% PVA in NMP (i.e. to adhere
71
strictly to the concentration reported by Ogura et al. [175-180,196-199]) was deposited
on each individual glass slide. The stock solution of PVA-NMP was previously generated
following a method described by van Zandvoort et al. [211] by dissolving a quantity of 6
mg of PVA, MW = 13,000-23,000 in 10 ml NMP and heating this mixture to 80°C under
continuous magnetic stirring until an optically clear solution was obtained. Attempts to
dissolve PVA at any molecular weight in NMP without heating the mixture at a
temperature of 80°C proved to be unsuccessful. Ogura et al. [175-180,196-199] had not
indicated that such heating was performed. The slides having the PVA-NMP mixture
were placed in an empty glass cell of the type depicted in Figure 15 and NMP-plasticized
films of PVA produced by evaporation of NMP under flowing pure argon at room
temperature for 24 hours. After the solvent evaporation, each glass slide was heated for
one hour at a temperature of 70°C to remove any remained moisture in PVA. Each slide
was then individually weighted and placed in a closed glass cell 2 mm above a
supersaturated solution of MgCl2 in water, in which a humidity of ~30% was established.
At all times, pure argon gas was flowed in the glass cell, in order to reproduce as close at
possible the humidity and gas flowing conditions encountered during the measurement
taken with emeraldine base-PVA composite films. At regular time intervals, the glass
slides were removed from the glass cell and weighted, the water absorption characteristic
being generated for each individual slide in the respective humidity level. Each glass
slide was then dried in an empty glass cell under argon gas flowing for 24 hours, dried at
70°C for one hour and weighted. The water absorption was then generated in other
humidity levels fixed by NaCl and LiCl supersaturated solutions (~60% and ~4.5%
relative humidity respectively).
72
2.2. Immersed Sensor
2.2.1. Conductive Polymer Preparation
Emeraldine base polyaniline MW = 5000 was purchased from Aldrich and treated
in 3% NH4OH aqueous solution for 8 hours. The solution was filtered through filter paper
No.1 (Whatmann Int.) and washed with several portions of methanol and deionized
water. The resulting precipitate was dried for 24 hours in vacuum oven at room
temperature and the dried polymer collected in a closed vial and stored at room
temperature in a dissecator. No heat treatment was performed on the emeraldine base
powder. Stock solution of 3% emeraldine base in NMP (i.e. a concentration
recommended by Zheng et al. [212]) was prepared by dissolving 0.3 mg emeraldine base
in 10 mL NMP, and stirring the solution with a magnetic stirrer on the top of a magnetic
plate at room temperature.
2.2.2. Polymer Film Deposition and Interrogation
The interdigitated electrode used for conductivity measurements was of the same
type as the sensor employed in the suspended sensor measurements (i.e. Figure 14), only
that the finger dimensions and interdigitated spacing differed. Each finger had a length (f)
of 3 mm, height (h) of 120 nm and an interdigitated spacing (d) of 15 µm to the adjacent
fingers. The interdigitated electrode cleaning and deposition procedures followed the
same methods previously described in the suspended sensor case.
73
2.2.3. Experimental Setup
A schematic of the experimental arrangement is presented in Figure 17. The
beaker was partially filled with pure deionized water. The interdigitated electrode
covered with the cast polymer film was connected with two alligator clips and the
working area immersed in water. The impedance measurement was performed on same
Impedance Gain Phase Analyzer as in the suspended sensor case for an identical
frequency sweep. In the immersed sensor case however, the detection of carbon dioxide
was assessed by exposing the sensor to two different conditions: immersing it in pure
water of measured pH ≅ 6.00 and in aqueous carbonic acid solutions of various pHs
lower than 6, with the lowest pH measured 3.95. During the impedance measurements,
the pH level of the aqueous solution was monitored inside the beaker with Accumet
Research AR 50 pH meter (0.01 accuracy).
Interdigitated
Electrode
Gas In
Electrical wires
Pure water
Beaker
Glass Tube
pH Meter
Figure 17. Experimental setup for the conductivity measurements in the immersed sensor case.
74
2.2.4. Measurement Conditions
The conductivity response as a function of pH of the protonating solution was
performed by immersing the interdigitated electrode in a beaker containing aqueous
hydrochloric solution of various pHs ranging from 6.00 to 0.25, with 0.25 pH units
decrement. Impedance measurements were taken in this configuration and the value of
the film total resistance extracted from the equivalent circuit fit.
In the immersed sensor case, the measured data is fitted with an equivalent circuit
displayed in Figure 16 or in Figure 18. The conductivity of the film was calculated by
inserting the total the resistance given by the equivalent circuit into equation (28).
RC C1
R1
C2
R2
C3
R3
Figure 18. Equivalent circuit schematic for the immersed sensor case.
The equivalent circuit displayed in Figure 16 (two series of parallel R-C elements)
was found suitable for fitting just a spectrum displaying a single semicircle, characteristic
to a single conduction mechanism activated. This single mechanism was active between
pH = 6.00 and pH = 5.03 exclusively for the sensor described in this project. The
equivalent circuit presenting a sequence of three R-C elements, displayed in Figure 18
was used for fitting an impedance spectrum composed of two semicircles, characteristic
to two conduction mechanisms based on electron-transfer. These two conduction
mechanisms corresponded to a range of pHs between 5.03 and 0.25 inclusively. In this
case, the total resistance of the polymer film was calculated by grouping the three
individual resistances into two groups of resistances according to the capacitance range of
75
the respective element. The high frequency capacitance range was 10-7 F to 5 × 10-9 F and
the low frequency capacitance range was 1 × 10-5 F to 10-7 F.
In the case of a spectrum displaying a single semicircle, in which case the
equivalent circuit displayed in Figure 16 was employed, R2 (also denominated Rlow
frequency) was taken as zero since there was no physical representation in the measured data
for such a resistor. Therefore, the total resistance of the sensor was calculated as a
summation of the two resistances given by the equivalent circuit, both of which were in
the high frequency range. Following the impedance measurements in hydrochloric acid,
the sensor was inserted for an hour in 3% aqueous solution of NH4OH in order to convert
the emeraldine salt to emeraldine base.
The assessment of the sensor response to carbon dioxide was accomplished by
performing impedance measurements with the interdigitated electrode containing
emeraldine thin film immersed in pure water or in pure water through which Ar + CO2
mixture was bubbled (i.e. carbonic acid solution). In subsequent experiments, the
impedance measurements were taken with the interdigitated electrode in pure water while
bubbling purified or unpurified argon gas, or in carbonic acid solution with no bubbling
performed during measurements. The purification of argon gas was performed through
capturing the gas resulted by bubbling the argon directly from a gas cylinder (Air Gas Co.
supplier, purity 99.9%) through a one molar solution of aqueous sodium hydroxide.
To evaluate the possibility that the carbonium ion induces complexation in imine
sites, an emeraldine base film from the same stock solution was deposited on a glass
slide, dried under vacuum at 45°C for 24 hours and then equilibrated in water for one
hour. In this time, the film peeled off the glass slide. An FTIR measurement was taken
76
for this film and then the film was reintroduced in water where pure CO2 was bubbled for
10 minutes, of measured pH = 3.95. The film was removed from water, tapped with Kim
wipe to absorb the moisture and an FTIR measurement was taken immediately. The time
spent by the film outside the carbonic acid solution before the FTIR spectrum was taken
was less than 15 seconds, the time necessary for mounting the film in the FTIR holder.
In order to assess whether the lack of complexation of imine sites is due to the
inherent NMP content in the film, an acid-base treatment was performed to the
emeraldine film and the respective FTIR spectra collected at different pHs generated by
HCl in water: starting from pH = 4 down to pH = 0.25 with 0.5 units decrement. The film
was re-equilibrated in 3% NH4OH for 30 minutes to convert the emeraldine salt to
emeraldine base. The film was then removed from the ammonium hydroxide solution,
rinsed with deionized water and equilibrated in water for 1 hour. An FTIR measurement
was subsequently taken for this film after removal from water and also for the same film
introduced in carbonic acid solution of measured pH = 3.95.
77
3. RESULTS AND DISCUSSION
3.1. Suspended Sensor
3.1.1. Experimental Results
The mechanism of CO2 detection proposed by Ogura et al. [175,177-180] is
displayed in Figure 19-adapted from reference [175]. The CO2 reacts with water from
PVA to create a carbonium ion (HCO3-) and protonates the polyaniline. Ogura reported
that as the CO2 partial pressure increases, the amount of protonation increases as well,
leading to a higher electrical conductivity of the polymer film.
2CO2 + 2H2O = 2H+ + 2HCO3-
N N Low CO2: N
H
N
H
N
H
HCO3-
⋅ + HCO3
- ⋅ +
High CO2: N
H
N
H
N
H
N
H
N
H
Figure 19. The CO2 detection mechanism.
An example of conductivity measurement using impedance spectroscopy is
presented in Figure 20, where the curve obtained is depicted as a depressed semicircle in
the Z’-Z’’ plot. The intercept of each semicircle with the horizontal axis represents the
real part of the impedance recorded by the Impedance Gain Phase Analyzer, in fact the
resistance of the emeraldine thin film in the respective measurement condition [209].
78
Figure 20. Impedance spectra recorded by Impedance Gain Phase Analyzer
The resistance variation during alternate flowing periods of argon and argon
containing 5% carbon dioxide is shown in Figure 21. The resistance variation was plotted
as a logarithm of the total resistance given by the equivalent circuit as a function of time.
Graphs displaying the response of the sensor for the other cycles employed are attached
in the appendix as Figure 21a-e. The stabilized resistances (i.e. plateau values), time and
humidity values for the suspended sensor investigated are displayed in Table 4 for all the
seven cycles (i.e. humidity levels) in which the sensor was tested.
-8000
-7000
-6000
-5000
-4000
-3000
-2000
-1000
0 0 2000 4000 6000 8000 10000 12000 14000 16000 18000
Z' (ohm)
Z'' (ohm)
Ar
Ar + 5% CO2
79
Table 4. Measured values for the suspended sensor.
Time
(min)
RH
(%)R (ΩΩΩΩ) LogR
Time
(min)
RH
(%)R (ΩΩΩΩ) LogR
1131 31.55 19543.4 4.291 1491 32.44 18793.2 4.274 3.84
1811 29.1 19678.9 4.294 3041 32.84 19230.9 4.284 2.283373 29.87 20230.2 4.306 4720 31.11 19319.7 4.286 4.504720 29.75 20137.2 4.304 5560 30.77 19498.4 4.29 3.176173 29.85 20511.6 4.312 7013 30.4 19408.9 4.288 5.38
2235 59.35 13867.6 4.142 3480 59.7 13001.7 4.114 6.244500 59.22 14060.5 4.148 6405 59.85 13213.0 4.121 6.037890 58.44 15205.5 4.182 9035 60.22 13931.6 4.144 8.3810835 58.15 15848.9 4.2 13805 59.72 14157.9 4.151 10.67
2715 29.95 15848.9 4.2 1380 31.75 14962.4 4.175 5.597005 29.74 16143.6 4.208 4065 30.87 15100.8 4.179 6.46
12870 30.14 16943.4 4.229 9855 31.68 15310.9 4.185 9.645730 3.67 21330.4 4.329 3180 4.26 18197.0 4.26 14.6910275 3.53 21379.6 4.33 7200 5.53 19543.4 4.291 8.5918690 3.34 21727.0 4.337 14490 4.55 19498.4 4.29 10.26
8235 1.15 17060.8 4.232 5280 1.31 15310.9 4.185 10.2615250 1.13 17864.9 4.252 12480 1.24 16595.9 4.22 7.107380 30.1 14157.9 4.151 4140 32.92 13273.9 4.123 6.2414505 30.44 14454.4 4.16 11265 32.02 13489.6 4.13 6.6721450 30.63 14927.9 4.174 17775 31.74 14028.1 4.147 6.03
4110 59.55 11749.0 4.07 1590 60.03 11091.7 4.045 5.597050 59.32 11995.0 4.079 5700 61.2 11091.7 4.045 7.53
∆∆∆∆R/R (%)
Salt
MgCl2
Argon Ar + 5% CO2
4
5
6
NaCl
LiCl
MgCl2
Cycle
1
2
3
7
MgCl2
NaCl
LiBr
80
4.24
4.25
4.26
4.27
4.28
4.29
4.30
4.31
4.32
0 1000 2000 3000 4000 5000 6000 7000
Time (min)
Log R
Ar + 5% CO
2
Ar + 5% CO2
Ar + 5% CO
2
Ar Ar Ar Ar Ar
Ar + 5% CO
2
Ar + 5% CO
2
Ar + 5% CO
2
Figure 21. Resistance of the composite film in Ar and Ar + 5% CO2 in a fixed humidity
level of RH ~ 30% (cycle 1 in Table 4).
All these graphs (Figure 21 and Figure 21a-e) show that a change in flowing gas
from Ar to Ar + 5% CO2 resulted in a decrease of the composite film resistance, which
according to Ogura et al. [175,177-180,198,199] could be explained by the inherent
change of the emeraldine component of the composite film from the insulating
emeraldine base to the conductive emeraldine salt following the reaction of carbon
dioxide with water molecules provided by the PVA matrix. These graphs reveal also the
undesirable drift of the sensor, in the way that the next measurement performed in the
same gas resulted in a higher stabilized resistance. The drift was constant in the sense of
resistance increasing with time, but the magnitude varied. The drift of the composite film
81
in the above-mentioned cycle 1 in Table 4 is presented in Figure 22. The drift was plotted
as the logarithm of the average resistance of each plateau as a function of time.
4.27
4.275
4.28
4.285
4.29
4.295
4.3
4.305
4.31
4.315
4.32
0 1000 2000 3000 4000 5000 6000 7000 8000
Time (min)
Log R
Argon
Ar + 5% CO2
Figure 22. Trend of the resistance response of the composite film in Ar and Ar + 5% CO2 at a fixed humidity level of ~ 30%.
The graphs of the resistance trend displayed by the sensor in the cases of other
cycles employed are attached in appendix as Figure 22a-e. The drift had a greater
magnitude in the case of argon gas flow than in argon containing 5% carbon dioxide.
Figure 21 together with Figure 21a in appendix show that immediately after the
gas switch from one to another the film resistance recorded by the sensor sharply
dropped. This event is visible only in these two figures since during the change, the line
was disconnected from the controller and another line from the second cylinder
connected, exposing in this way the glass cell with the room ambient. For the cycles 3 to
82
7 inclusively in Table 4, a second flow controller was calibrated and connected through a
T switch, allowing identical flow rates to be passed through the test chamber, from the
two cylinders. The event of sharp resistance drop was not recorded any more in the last
five cycles.
Figure 23 displays the overall trend, for all the humidity conditions in which the
sensor was tested. The overall trend is plotted as logarithm of the average resistance of
each plateau as a function of relative humidity in all four humidity conditions: 1.5%,
4.5%, 30% and 60%. Figure 23 reveals that the resistance of the sensor does not stabilize
to the same set of values when reverted to a humidity level already employed.
4.00
4.05
4.10
4.15
4.20
4.25
4.30
4.35
0 10 20 30 40 50 60 70
Relative Humidity (%)
Log R Cycle 6
Cycle 3
Cycle 1
Cycle 2
Cycle 7
Cycle 4
Cycle 5
Argon
Ar + 5% CO2
Figure 23. Logarithm of resistance versus relative humidity in argon and Ar + 5% CO2
for the composite polymeric film. The film was tested more than once in both magnesium and sodium chloride respectively. Each particular cycle was circled for the ease of
visualization.
83
Figure 24 displays the resistance difference between argon and argon containing
5% carbon dioxide. The resistance difference between argon and argon containing 5%
carbon dioxide is plotted as a ratio of resistance difference in the two gases to the
resistance displayed in argon, as a function of relative humidity.
The graph shows that the magnitude of response of the sensor is lower in the
humidity level fixed near 30%. In this case, the ∆R/R response varies between 2.2% and
9.64%. In a humidity level of ~60%, the magnitude of response is placed between 5.59%
to 10.67%. For a low relative humidity level fixed by LiCl and LiBr in the interval 1.1%
to 4.5%, the magnitude of response is placed between 7.10% to 14.69%.
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
0 10 20 30 40 50 60 70
Relative Humidity (%)
∆∆ ∆∆R/R (%)
% Resistance Change
Figure 24. Resistance difference between Ar and Ar + 5% CO2 (logarithmic scale).
A useful piece of information is the time necessary for the sensor to reach 90% of
the response. An example of such variation is provided in Figure 25 for a humidity level
84
of ~30%. The 90% time response (t90) was calculated as the time necessary for the
sensor to reach 90% resistance between two adjacent plateaus (i.e. the one corresponding
to argon gas and the other to Ar + 5% CO2). This value was plotted as a function of
number of gas reversal from argon to argon containing 5% carbon dioxide. Figure 25a in
appendix shows an example of calculation of t90.
0
200
400
600
800
1000
1200
1400
0 1 2 3 4 5Gas Cycle
T90 (min)
Argon
Ar + 5% CO2
Figure 25. Response time (t90) reached by the composite film for ~60% RH (cycle 2 in Table 4).
The inspection of the 90% response time data in Figure 25 and Figure 25a in
appendix reveals that the change of gas from pure argon to argon containing 5% carbon
dioxide resulted in a faster response (a shorter time necessary for the sensor to reach the
plateau), even though there were exceptions in almost every humidity level. This fact
complements the observation that the sensor output drifted. Otherwise, the faster
85
response during the exposure to carbon dioxide was not a surprise, this behavior being
also observed by Ogura and his coworkers [175,177-180,198,199]. This phenomenon was
ascribed by Ogura to the slow removal of the weakly bound water in the case of inert gas
flow, which would prevent the resistance of the sensor to recover quickly the initial
value. Travers et al. and Alix et al. [191,193] generated a similar observation pertinent to
the existence of weakly and strongly bound water molecules in the polyaniline chains.
However, the experimental method followed by Ogura et al. was different from the one
in the present work, fact that permitted them to assign the behavior on the removal of the
weakly bound water molecules from the composite polymer chain. The group used a
vacuum pump to evacuate the chamber, allowing in a following step the admission of a
specific amount of water vapor with carbon dioxide. Then the pressure in the glass cell
was raised to one atm by admitting nitrogen and the measurement taken. This explanation
offered by Ogura is not valid for the present work, since in the current project the
polymer was not cycled between dry and wet states. Instead, the relative humidity was
fixed by supersaturated salt solutions and just the gas flowed into the glass cell was
changed between Ar to Ar + 5% CO2. However, even in this work, the relative humidity
varied between two gas flow conditions, in all the cases the relative humidity displayed in
argon gas being higher up to 1.5% units than in the case of Ar + 5% CO2. This difference
was not due to the calibration of the flow controller since the flow rate was always the
same (e.g. for the cycles 1 and 2 the same controller was employed, whereas in the cycles
3 to 7 two different controllers were utilized). In addition, the humidity sensor detected
all the time a higher relative humidity when the gas from the cylinder containing 5%
carbon dioxide was flowed through the glass cell.
86
The similarity of an apparently faster response time in the case of an exposure to a
higher relative humidity, between this work and Angelopoulos [194], as well as the very
slow response of the sensor (i.e. t90 that reached even 24 hours in this work when the
sensor was exposed to just a percent variation in relative humidity and t90 that reached 10
hours in Angelopoulos et al. work [194] for a sensor cycled between dry state and ~17%
relative humidity), suggest that in fact the sensor in this work detected just humidity
variation.
Therefore, if in the current project the sensor did not detect a change in gas
composition but just a humidity variation, the above models of dependence of emeraldine
on relative humidity variation described by Angelopoulos et al. [194] and Travers et al.
[191,192] could be well suited for an explanation of the poor results of the sensor in this
work. In the same time, the same model, would explain Ogura’s 1-1.5 order of magnitude
response of their sensor. This magnitude response was detected by Travers et al. [191] as
well as other authors (see reference 6 in 191). Salanek et al. obtained several orders of
magnitude response in conductivity for a sample of emeraldine base with a specific
fabrication history cycled between dry and wet states.
3.1.2. Sensor Performance Evaluation
In a similar work, Ogura et al. [177] reported a sensor able to detect CO2
concentrations ranging from 50 to 10000 ppm (i.e. 0.005% up to 1% CO2).
In the present work, the composite film was only tested for 5% CO2
concentration. Since the decrease in the resistance of the film when changed from pure
argon flowing conditions to argon containing 5% carbon dioxide, was placed in the range
87
of 2.2% to 14.6%, the composite film was not tested for carbon dioxide concentrations
lower than 5%.
The purpose of the sensor is an important consideration in judging the response
time. If the sensor is designated to monitor the food quality in a regular food package, a
slow response time is acceptable. However, if the application involves sequential
measurements, a fast response time in the range of seconds is required. Preliminary
results showed that the sensor reached 90% of the response in a time of hours and the
response was in most of the cases, faster when Ar + 5%CO2 flowed through the glass
chamber. At this stage of development, it is not clear which represented the factors that
triggered a slow response time of the sensor and why the sensor response was apparently
faster upon carbon dioxide exposure.
Preliminary results also showed that the sensor output was not stable, but drifted
in time. The drift resulted in an increase of the composite polymer film resistance with
time, but the values of the dynamic range fluctuated between the two conditions (Ar and
Ar + 5% CO2) and between different humidity cycles employed.
3.1.3. Comparative Results
Compared to the results reported by Ogura et al. [175,177] the sensor developed
in the current project has less appealing features.
1) The 90% response of the sensor (t90) in the current project (in the range of hours,
up to more than 24 hours) is orders of magnitude larger than the respective values
reported by Ogura et al. [175,177] (0.9 seconds for CO2 flowing and 7.2 seconds for
nitrogen).
88
2) The dynamic range of the sensor is much smaller (2.2 to 14.6% change in
conductivity) between the two gases. In addition, the dynamic range is not stable, being
arbitrarily smaller or greater between two similar flowing conditions and between
different humidity levels employed. The sensor reported by Ogura et al. [175,177]
presented a 1 to 1.5 order of magnitude increase in conductivity in response to carbon
dioxide exposure.
3) The stability of the sensor is severely disrupted by drift. The sensor reported by
Ogura et al. [175,177] was stable and presented a minimal drift over 35 days of testing.
The drift and variation of the dynamic range, slow response time and instability
made the sensor developed in present work unsuitable for carbon dioxide monitoring
applications. Given these unacceptable results, the focus of the subsequent development
was aimed towards designing and testing of polyaniline films responsive to a
concentration of CO2 lower than 5%, and the quantification of the minimum
concentration for a reliable response. In fact, the limited dynamic range of the sensor, the
poor stability and slow response time provided suspicion that the sensor did not detect
carbon dioxide at all, but in fact responded to humidity variations. The setup with the
sensor immersed in water (i.e. the one to be employed in the following development of
the project), along with the respective results, will clarify this issue. In the immersed
sensor case, the amount of water to which the emeraldine base is exposed during carbon
dioxide in and out stages will not be an influential factor on conductivity since the sensor
being inserted in water is exposed to the same water level at all times.
89
3.1.4. Materials Characterization
Polymer characterization helps describing and understanding the structural and
chemical modifications of the polymer film that accompany the conductivity changes. An
extremely necessary piece of information is the determination of a change in intensity of
the benzenoid to quinoid bands which provides information about the amount of
protonation of imine nitrogen sites in the polymer backbone and therefore about the
expected polymer conductivity.
3.1.4.1. FTIR Spectroscopy Characterization
A comparison of the FTIR spectra among emeraldine base powder not heat-
treated and samples heat-treated for 1 hour at various temperatures in helium gas
atmosphere is presented in Figure 26, with all the characteristic peaks marked on the
graph. The spectra of samples heat-treated to temperatures intermediate between these
two extremes are inserted in appendix as Figure 26a-e for a better visualization of the
peaks.
90
0
50
100
150
200
40080012001600200024002800320036004000
Wavenumber (cm-1)
Transm
ittance (a.u.)
3365 cm-1 3284 cm
-1
1596 cm-1
1497 cm-1
1299 cm-11145-1164 cm
-1
1040 cm-1
3404 cm-1
3375 cm-1
No heat
treatment
200°°°°C treated
250°°°°C treated
300°°°°C treated
365°°°°C treated
Figure 26. Comparison between FTIR spectra of emeraldine base without heat treatment and heat treated at various temperatures for 1 hour in helium atmosphere.
The peaks revealed by the FTIR spectra are displayed in Table 5 along with their
respective vibrational mode already assigned in the literature [170,175,177,213-219].
The analysis of FTIR spectra reveals that emeraldine base suffers a gradual
transformation as a result of heat-treatment over 250ºC. In fact, the analysis of the FTIR
peaks demonstrates that subsequent to the treatment at temperatures higher than 250°C,
the emeraldine powder becomes a mixture of leucoemeraldine and cross-linked
emeraldine base.
91
Table 5. Characteristic peaks and vibrational modes assigned to emeraldine base
Peak (cm-1) Vibrational Mode Reference
3365 Free N-H stretching 212,213,216,217
3285 Interchain H bonded N-H 212,213,216,217
1590 C=C stretching-quinoid rings 170,175,177,212-218
1495 C=C stretching-benzenoid rings 170,175,177,212-218
1295 C-N stretching-aromatic amine 170,175,177,212,213,216-218
1145-1165 N=Q=N stretching 170,177,212-218
1040 Symmetric SO2 stretching in 177
Scherr et al. [184] reported that emeraldine base cross-links if treated for 4 hours
at 300°C and proposed that the cross-linking process is chemical in nature and occurs
through a linkage between imine nitrogens that generate a phenazine-type structure. A
schematic of the cross-linking proposed by Scherr et al. is presented in Figure 27-
adapted from reference [184].
92
T> 250ºC
N
N
N
H N
H N
N
N
H N
H
N
N
N
H N
H N
N
N
H N
H
N
N N
H N
H
N
H
N
H N
H
N
H
N
H
N
H N
H
N
H
N
H N
H
N
N
Figure 27. Schematic of the crosslinking process.
This behavior of cross-linked emeraldine was confirmed by Oh et al. [220] in a
study aimed to identify the effects of cross-linking induced by both temperature and
stretching in emeraldine base films. The latter report also revealed that the degree of
cross-linking and the molecular weight of polyaniline employed have little effect on the
conductivity in cross-linked polyaniline films. Chen et al. [215] also confirmed the
chemical nature of cross-linking process occurring at temperatures of 150-300°C,
showing that the process is responsible for the diminishing in the intensity of the peak
centered at ~1160 cm-1 and attributed to N=Q=N stretching type vibration. In a following
report, Milton and Monkman [221] reported that chemical cross-linking process rendered
by treating the sample at temperatures greater than 180°C is responsible for the sample
93
hardening and insolubility in NMP. The latter group suggested that physical cross-links
(chain entanglements) may also be present in a cross-linked sample. In a more recent
report, Ding et al. [214] confirmed Milton and Monkman study showing that mainly the
chemical cross-links between imine nitrogen sites accompanies the change in
morphology of emeraldine base due to the thermal treatment performed in the range of
185-350°C. Ding et al. based their argument on the evidence of transformation of the
multiple peaks centered at ~1160 cm-1 (i.e. peaks which are characteristic to a more
ordered, resembling a partially crystalline structure) into a single peak of much lower
intensity (i.e. a single peak representative to a fully amorphous structure. The chemical
nature of cross-linking process, occurring through bonding between imine nitrogens, is
consistent with the FTIR spectra in Figure 26. The multiple peaks having an infrared
absorption in the range of ~1144-1165 cm-1 gradually diminish in intensity with
increasing temperature and finally vanish for an emeraldine powder treated at a
temperature of 365°C. In addition to the FTIR analysis, Ding et al. [214], Chen et al
[215], Scherr et al. [184] and Monkman et al. [221] reported that once the emeraldine
powder became cross-linked at high temperature, the powder could not be processed in
appropriate solvents (i.e. DMSO or NMP) and strong acids (i.e. H2SO4). In this work, the
powder heat treated at temperatures exceeding 250°C it is never completely soluble in
NMP. The higher the heat-treatment temperature, the lower the solubility of the powder
and less intense the blue color of the resulting solution. At heat temperatures over 300°C,
the color of the solution obtained after the dissolution of the respective powder is no
longer blue, (the color which appears after the dissolution of emeraldine base in NMP)
but yellowish-brown, characteristic to the leucoemeraldine-NMP solution. In about 24
94
hours time, a residue of undissolved polymer sediments at the bottom of the flask. The
dissolved polymer is actually the leucoemeraldine base (as determined by UV-visible
spectrophotometry), whereas the undissolved one is presumably in the form of cross-
linked emeraldine base.
The spectrum of emeraldine base for which no heat treatment was performed (i.e.
the lower curve in Figure 26) displays two bands associated with N-H stretching
vibrations: a broad band centered at ~3285 cm-1 associated with hydrogen bonded N-H
stretching vibrations and a minor band centered at ~3365 cm-1 associated with non
hydrogen bonded free N-H type stretching vibrations. These two values are in good
agreement with the published values of ~3290 and ~3383 cm-1 respectively
[212,213,216,217]. In contrast, the spectrum of leucoemeraldine base contains just the
band at ~3383 cm-1, which is characteristic to free, non-hydrogen bonded N-H stretching
vibrations [212,213,216,217].
The sequence of spectra taken for emeraldine base powders after heat-treatment at
temperatures higher than 250°C demonstrate the disappearance of the hydrogen bonded
N-H stretching vibration band centered at ~3285 cm-1, and the persistence of only one
peak, at ~3365 cm-1. In addition, the intensity of both quinoid and benzenoid bands is
severely reduced, and the peaks broadened. The latter facts, together with the
disappearance of the hydrogen bonded N-H stretching vibration, indicate that at
temperatures exceeding 250°C, the emeraldine base undergoes a chemical
transformation, being reduced to leucoemeraldine base. A similar discovery was reported
by Mathew et al. [217,219].
95
Figure 26 shows that the peak centered at 1040 cm-1 due to the SO2 symmetrical
stretching vibration (i.e. the peak that shows the presence of TSA residue in the
emeraldine chains) is severely reduced during the heat treatment until it finally disappears
at treatments performed at temperatures exceeding 300°C. The removal of the TSA
residue from the polymer chains by heating the polymer at 380°C represents the purpose
of heat treatment performed by Ogura et al. [177] and is probably the only desirable
effect of the high temperature exposure of the polymer.
3.1.4.2. Differential Scanning Calorimetry Characterization
The cross-linking of emeraldine base as a result of heat treatment can be
unequivocally demonstrated through a DSC experiment similar to the one performed by
Wei et al. [187]. The analyzed sample was kept isothermal for 30 minutes at 100°C and
then ramped at 20°C/min to 400°C. After cooling down to the room temperature, a
second run of the experiment was restarted in the same described conditions. Due to the
isothermal held at 100°C, the temperature is displayed on graph starting with 115°C.
The DSC thermogram in Figure 28 shows that the emeraldine powder heated at
400ºC presents an exothermic peak centered at about 225ºC (~250ºC in Wei et al.
experiment). The DSC thermogram taken right after cooling down the sample to the room
temperature displays no exothermic peak, indicating that an irreversible process had
taken place during the first run. A thermo gravimetric analysis (TGA) of the emeraldine
polymer shows no major loss in the interval 200 to 300ºC, a fact revealed by many
authors [150,159-161,164,165,169-171,173,184,187,190,212,213], therefore the thermal
process that appears at ~225ºC can be attributed to the crosslinking of the polymer chains
96
as suggested by Wei et al. [197] and also other authors [184,213,215]. Emeraldine base
should display a small endothermic peak in the range 200 to 250ºC due to the glass
transition that takes place at a temperature in this interval [184,187,213,215], but this
endothermic peak is offset by the large exothermic peak due to crosslinking of the
polymer chain. With this respect, DSC is not a suitable technique to visualize the glass
transition in emeraldine base, but the missing endothermic peak in the DSC spectrum
suggests that the glass transition occurs in the vicinity of crosslinking temperature.
-12
-10
-8
-6
-4
-2
0
2
4
6
115 165 215 265 315 365
Temperature (C)
Heat Flow (W/g)
100 ºC - 400 ºC - first run
100 ºC - 400 ºC - second run
Figure 28. DSC thermogram of emeraldine base held isothermal for 30 minutes at 100ºC and ramped up 20ºC/min to 400ºC.
3.1.4.3. X-Ray Diffraction Characterization
The cross-linking phenomenon occurring at temperatures exceeding 250ºC is also
shown in the X-ray diffraction pattern of the powder samples treated at various
97
temperatures for one hour in helium atmospheres. The peaks are indexed for the
emeraldine base class I (EB-I) crystal structure, according to Jozefowicz et al. [206]. A
comparison of XRD spectra between emeraldine base not heat treated and samples
treated at various temperatures is presented in Figure 29. A calculation of the full width at
half maxima (FWHM) as well as the ratio of half width over peak height (HW/H) is
provided in Table 6. An example of calculating the FWHM and HW/H for an XRD
spectrum of a sample heat treated at 100ºC for one hour in helium atmosphere is inserted
in appendix (Figure 29a).
0
500
1000
1500
2000
2500
3000
3500
0 5 10 15 20 25 30 35 40 45 50
2 θθθθ (deg)
Intensity (a.u.)
110
111 200
211
No Heat Treatment
225 ºC
245 ºC
365 ºC
Figure 29. X-ray diffraction of emeraldine base treated at various temperatures.
98
Table 6. Full width at half maxima and peak half width over peak height for emeraldine base samples heat-treated at various temperatures.
Heat Temperature (ºC) FWHM HW/H
No Heat Treatment 2.3 0.58
100 2.2 0.57
145 2.1 0.5
175 1.9 0.48
205 1.7 0.49
225 1.6 0.47
245 1.8 0.59
275 1.9 0.64
300 2.5 0.68
365 14.1 0.98
If crystallinity is present in polyaniline powder samples, then the main crystalline
peak should appear at an angle 2θ ≅ 19.5 [212,222-224]. The crystalline structure in
emeraldine base is not expected if the powder was originally prepared in the conductive
salt (emeraldine salt) form and then converted to its corresponding insulating (emeraldine
base) form through treatment with a strong base, which was in fact the case of the powder
prepared in this work, following Ogura’s method (i.e. EB-class I). According to Pouget et
al. [222,223] and Jozefowicz et al. [224] crystallinity should be expected just in the case
of emeraldine base powders directly synthesized in the base form (i.e. EB-class II), or
respectively, emeraldine salts directly synthesized in the salt form (i.e. ES-class II).
99
Studying the structure of polypyrrole and the effect of dopant anion on
crystallinity of the polymer, Warren et al. [225] found that the ratio of the peak half width
to peak height (HW/H) of the X-rays diffraction peak reflects order in the polymer
backbone. Thus, the smaller the value of the HW/H ratio, the higher the order. Studying
the properties of a polyaniline film, Wei et al. [187] detected the same relationship
between HW/H and order, and found that the value of the ratio HW/H decreased with
increasing temperature, reached a minimum and then increased again as the temperature
further increased. In their study, the X-Ray diffraction of emeraldine film (which
inherently contains ~15% NMP that could not be removed by dynamic pumping)
presented the lowest ratio HW/H at ~150°C, which was exactly in the vicinity of glass
transition temperature for NMP plasticized emeraldine film. An NMP-free emeraldine
base film, however, employed in the same study (prepared by performing a chemical
treatment of the film to remove the residual NMP) resulted in a glass transition
temperature of ~250°C and a maximum order of the polymer given by lowest HW/H ratio
of the X-Ray diffraction peaks corresponding to that temperature.
Considering these, the values obtained in this work are in good agreement with
the literature [187] since in this case the glass transition temperature fell in the interval
200 to 250°C, as proved by the DSC experiment. In Wei et al. experiment [187] the
calculated HW/H ratio of EB film was 0.70 for a sample treated for one hour at 40°C,
reached a minimum of 0.49 when treated for 1 hour at 150°C (i.e. the glass transition
temperature) and then increased with further raise of heat temperature up to a value of
0.91 for a heat treatment of 30 minutes at 280°C.
100
However, the full width at half maxima (FWHM) is a more exact parameter for
analysis of the degree of order in emeraldine base, especially in cases (as in this work) in
which the EB-I polymer displays some finite degree of crystallinity. This is because the
FWHM criterion does not have to take into account the deconvolution of the overlapping
peaks when calculating the base width. The FWHM values displayed in Table 6 are also
in agreement with the pattern reported by Wei et al. [187], the polymer displaying a
maximum order at a temperature of ~225°C, therefore in the vicinity of its glass
transition.
3.1.4.4. UV-Visible Spectrophotometry Characterization
UV-Visible spectrophotometry supports the argument introduced by Matthew et
al. [217,219] regarding the transformation of emeraldine base in leucoemeraldine base
during the high temperature treatment. The collected spectra taken for emeraldine base
samples treated at various temperatures are presented on the same graph in Figure 30.
The displays the two peaks of interest: a peak centered at ~328 nm due to a π-π*
absorption of benzenoid rings and a peak centered at ~630 nm due to exciton absorption
of quinoid rings, in good agreement with values reported in the literature [154,166-
173,183,217,226-229].
The emeraldine base to leucoemeraldine base transformation after treatment at
temperatures exceeding 250ºC is proved by the gradual disappearance of the peak at
centered at ~628 nm and the persistence instead of only the peak centered at ~328 nm.
The sequence of the spectra show that with an increase in the heat treatment of the
emeraldine powder, the polymer slowly transforms into its corresponding
101
leucoemeraldine base, the latter being shown in the literature as having characteristic just
the peak centered at ~328 nm [151,154,212,215,217,227-229].
0
0.5
1
1.5
2
2.5
3
200 300 400 500 600 700 800Wavelength (nm)
Absorbance (a.u.)
100C
150C
200C
250C
300C
365C
328 nm-Benzenoid ring absorption
630 nm-Quinoid ring absorption
Figure 30. UV-Visible spectrum of emeraldine base samples treated for one hour at
various temperatures in helium atmosphere.
3.1.4.5. Materials Characterization Discussion
All the material characterization techniques show that the heat-treatment process
(employed by Ogura et al. [175-180,197,199] in their development of the emeraldine
base CO2 sensor) represents an unnecessary step. Unfortunately, the removal of the
toluene sulfonic acid residue from the polymer chain is more than offset by the other
undesirable transformations that concomitantly occur: the partial crosslinking of the
polymer and the transformation of part of emeraldine base to leucoemeraldine base.
These events might explain why the dynamic range of the sensor is very narrow.
102
The leucoemeraldine base is processable in NMP but the film deposited cannot be
protonated and therefore it does not display an increase in conductivity in the presence of
any protonating agent. Moreover, the cross-linked emeraldine base was reported in the
literature to be virtually insoluble in NMP, therefore the proportion of true emeraldine
base chains in the deposited polymer film is expected to be very low compared to the
leucoemeraldine base chains.
The choice of toluene sulphonic acid as a protonating agent for emeraldine base is
in fact not a good one. Studying the properties of methyl derivated emeraldine (poly o-
toluidine), Jozefowicz et al. [230] showed that the existence of the CH3 groups in the
emeraldine chain leads to 5% increase in interchain separation and reduced interchain
order. This slight increase is responsible for about three orders of magnitude decrease in
room temperature conductivity of POT-ES compared to ES-I. X-ray data supports the
theory that methyl groups lead to one dimensional electron localization in the poly o-
toluidine chains, due to higher interchain separation, interchain zigzag and possible
disorder generated by randomness in placement of the methyl groups. This random
distribution of CH3 units would further generate disorder in the ring tilt angle and
position of the chlorine ions. Compared to POT-ES, the emeraldine salt, ES-I displays
three dimensional electron delocalization within crystalline regions. MacDiarmid and
Epstein [182] also support the theory of lower conductivity induced by the nature of the
protonating agent by showing that bulky groups like dimethyl-sulphate (i.e. (CH3)2SO4)
induce lower Pauli susceptibility and two order of magnitude lower conductivity of the
resulting emeraldine salt polymer compared to the emeraldine salt fabricated with HCl as
the protonating agent. This difference in behavior is explained by MacDiarmid and
103
Epstein [182] by the presence of bulky methyl groups which would disrupt symmetry and
increase localization of charge due to interchain separation. The authors consider that the
charge localization is induced when the dimethyl-sulphate group is not attached to each
nitrogen atom in the polymer, making possible in this way an uneven distribution of
charge along the polymer chain.
Toluene sulphonic acid (C7H8SO3) has an even bulkier functional group, the
effect of doping of emeraldine with TSA leading to charge localization and implicitly a
lower conductivity in the protonated emeraldine salt. It is therefore not clear why Ogura
et al. [177] chose this protonating agent in their study. Moreover, the respective group
finishes the fabrication procedure by a heat treatment in order to eliminate the residue
from the emeraldine base chain. It would have been much easier to employ hydrochloric
acid as in the classical method of fabrication proposed by MacDiarmid [160] and
followed by most other groups over the years [144,148,149,152,156,158,159,169,181],
which would not have necessitated any heat treatment, and therefore any ambiguity in the
nature of the resulting material. MacDiarmid and Epstein [182] in fact suggested that
hydrochloric acid is the most appropriate agent in polyaniline synthesis, since the residue
being volatile, is easily removed under vacuum, unless one tries to study per se the
properties of a specific agent-doped polyaniline, like TSA for example.
3.1.5. Possible Explanations for Poor Performance
3.1.5.1. Tautomerism
Tautomerism or hydrogen migration effect was discovered by Shimano and
MacDiarmid [227,228] in a study aimed to elucidate the unexpected photoluminescence
104
properties of some emeraldine base films previously reported by Chen et al. [231] and
Wan et al. [232]. Shimano and MacDiarmid proved that this phenomenon is activated by
water absorption in the polymer film during the film casting or solution preparation
process. Water absorption in NMP during the film cast and solvent evaporation stages
leads to hydrogen migration in emeraldine chains if the relative humidity falls in the
interval 43 and 57% or the NMP in which the emeraldine was dissolved previously
absorbed between 9 and 18 vol.% H2O. If at least one of the above conditions was
fulfilled, the induced photoluminescence persists in the dry emeraldine film irrespective
of the surrounding humidity level [227,228]. The existence of photoluminescence in cast
emeraldine base films demonstrates that the films actually contain extended sequences of
leucoemeraldine base, since only leucoemeraldine displays photoluminescence
[227,228,231,232]. Tautomerism makes therefore the EB chains transform in sequences
of EB-LB-PB. Only emeraldine base can be protonated, so the tautomerism leads to the
generation of a polymer chains composed of segments than can be made conductive
interrupted by segments than can never be made conductive (leucoemeraldine and
pernigraniline). A schematic of the process is presented in Figure 31-adapted from
reference [227].
Only the absorption of water between 9 and 18 vol.% in NMP could be responsible
for tautomerism in this work, since the relative laboratory humidity measured with the
humidity sensor was always in the range 34 to 37%, therefore outside the range suggested
by Shimano and MacDiarmid [227,228] to be responsible for inducing tautomerism.
105
Four Benzenoid Units Two Quinoid Units
Solution
2 Benzenoid Units 2 Benzenoid Units 1 Quinoid Unit 1 Quinoid Unit
Tautomerism
Leucoemeraldine base
(Extended Sequence)
Emeraldine base
(Extended Sequence)
Pernigraniline base
(Extended Sequence)
Cast Dried Film
N
H
N
H
N
H
N
H
N N N N
N
H
N
H
N
H
N
H
N N N N
Figure 31. Schematic of the tautomerism process.
3.1.5.2. Limited Protonation in Emeraldine Base
MacDiarmid and co-workers [160-162], followed by Epstein et al. [233] and
Jozefowicz et al. [234] showed that very little or virtually no protonation of emeraldine
base occurs if the pH of the protonating bath is greater than 4. A supporting argument is
presented in Figure 32-adapted from references [160,161], with the original experimental
points collected by respective authors displayed on graph. The use of DataThief software
helped generating the graph line close to the originally published line. The graph shows
that emeraldine base is not protonated until the pH of the doping hydrochloric solution
reaches the value of 4. The sample is in fact not significantly protonated even at a pH of
106
2, whereas a maximum protonation level is reached when the pH of the bath approaches
zero. Reiss [235,236] considered that the method of calculating the doping percentage
followed by authors in references [160] and [161] (i.e. chlorine uptake) was not free of
experimental errors, either from a failure to reach the full equilibrium in the protonating
bath, or from shifting of that equilibrium during the drying process of the sample. It is
therefore reasonably to consider that higher errors occur at high pH values, where the
chlorine uptake is smallest.
-101234567
pH
Doping Percentage (%
)
0
10
20
30
40
50
Figure 32. Doping percentage as a function of pH of protonating solution for the class I of emeraldine base.
To calculate the expected pH of a solution in which 100% carbon dioxide was
bubbled, the series of equations (3) to (7) were used. Therefore,
K1 = )]([
][][
2
3
aqCO
HCOH−+ ×
= 4.47 × 10-7 M/atm (7)
107
but the proton and bicarbonate concentrations are equal in carbonic acid solution more
concentrated than 10-5 M [37] (i.e. pCO2 above 355 ppm), given:
K1 = )]([
][
2
2
aqCO
H +
(29)
Therefore in the above equation, [CO2(aq)] is the only unknown value for the pH
calculation. To solve for this, the above equation (2) is employed:
CO2(aq) = pCO2 × KH (2)
where KH = 30 ×10-3 M/atm is Henry’s constant.
Finally, [H+]2 = (K1 ×KH × pCO2) and pH = - log ([H+])
For water in equilibrium with 100% rich CO2 atmosphere (i.e. pCO2 = 1),
plugging the values of K1 and KH in the above equation gives the pH = 3.936. At this pH
level induced by bubbling 100% carbon dioxide in water, following the graph published
by MacDiarmid [160,161] and reproduced in Figure 32, EB is virtually not protonated
and therefore does not display any perceptible increase in conductivity.
3.1.5.3. Limited Amount of Water in PVA Matrix
Other possible explanation for such large discrepancies between the results in this
work and Ogura et al. may be due to the amount of water incorporated by poly(vinyl
alcohol). A graph showing the amount of water absorbed measured by mass gain is
presented in Figure 33.
108
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0 1000 2000 3000 4000 5000 6000
Time (min)
Water Absorbed (%)
RH ~ 5%
RH ~ 30%
RH ~ 60%
Figure 33. Percent water absorbed in poly (vinyl alcohol)-NMP plasticized film, at various humidity levels.
The amount of water absorbed by PVA polymer in powder form was reported to be as
high as 20% by weight at 80% relative humidity [237]. However, the plasticized PVA
films were reported to absorb less amount of water than PVA powder. Studying the water
uptake in plasticized PVA and cellophane films, Creutz and Wilson [238] showed that a
relative humidity of 57.5% is responsible for 5.1 % by weight water uptake in PVA films.
Therefore, the water uptake level measured thermogravimetrically in this work is a factor
of 3.5 smaller than the value reported by Creutz and Wilson [238]. The limited quantity
of water inside the composite polymer (emeraldine base-PVA) chains may explain the
poor performance of the composite polymer employed for carbon dioxide detection,
109
though the amount of water necessary for the carbon dioxide hydrolysis reactions is
arguable.
3.2. Immersed Sensor
3.2.1. Experimental Results
Due to the deleterious effect of the presence of TSA residue within the polymer
(e.g. see the above described electron localization leading to a lower potential
conductivity) as well as the undesired transformations induced in emeraldine by the
thermal process to remove the sulphonate residues, the use of TSA-doped emeraldine was
discontinued and emeraldine hydrochloride was employed instead for the rest of the
project.
The study revealing the limited amount of water absorbed in the PVA matrix
generated the idea to immerse the sensor in water. In this arrangement, carbon dioxide
was bubbled through the water. Since water became in this experimental set-up
abundantly provided to the emeraldine base polymer chains, the presence of PVA matrix
as a water pool was not necessary any more. Therefore, an emeraldine base in NMP film
alone was deposited on a 15 micrometers spaced interdigitated electrode.
A graph showing the two impedance spectra, corresponding to a measurement in
pure water of pH = 6 and carbonic acid solution of pH = 4.4 is displayed in Figure 34.
During the impedance measurements in this configuration, the pH of the aqueous
solutions in which the sensor was inserted was continuously monitored with a pH meter
with no buffer employed to fix the pH.
110
Figure 34. Impedance spectra in pure water and water where pure CO2 was bubbled.
The above graph reveals that a change of medium from pure water of measured
pH ≅ 6.00 to carbonic acid solution of measured pH ≅ 4.40, does not cause any change in
total resistance of the thin emeraldine film since the two measurements present the same
intercept on the horizontal axis [209]. A conductivity measurement performed with a DC
technique would have measured only the total resistance. Impedance spectroscopy on the
other hand is a powerful technique because it allows one to separate contributions from
different conduction processes and determine the frequency interval over which each of
them is dominant.
A spectrum collected in pure water shows only one mechanism active, with a
characteristic peak frequency of 6385 Hz. When carbon dioxide is bubbled and
subsequently carbonic acid dissociates producing protons and carbonium ions, the pH
drops to ≅ 4.40 and the emeraldine film senses this new medium. Apart from the original
-9000
-6000
-3000
0 0 3000 6000 9000 12000 15000
Z' (ohm)
Z'' (ohm)
f = 6384.8 Hz
f = 6384.8 Hz
f = 32 Hz
f = 201.91 Hz
H 2 O in air, pH = 6.00 H 2 O in CO 2 , pH = 4.4
111
mechanism, which is still present, and display the same peak frequency of 6385 Hz, an
additional mechanism having a characteristic peak frequency of 32 Hz becomes active at
a frequency of 202 Hz.
In the following part of the project, attempts are made to characterize and
interpret this behavior not yet reported in the literature.
3.2.1.1. Conductivity Response in Hydrochloric Acid Solutions
Figure 35 shows the resistance response of emeraldine base thin film sensor, in
hydrochloric acid of various pHs. For each specific pH level the resistance displayed on
graph represents the value provided by the equivalent circuit fit.
0123456pH
Resistance (ohm)
100
101
102
103
104
105
Figure 35. Total resistance of emeraldine base thin film vs. pH in aqueous HCl solutions.
112
A comparison between the conductivity response of the film calculated with
equation (28) as a function of pH and the one published by MacDiarmid (i.e. adapted
from the original reference, with the original data points preserved by DataThief
software) is presented in Figure 36. However, the two polymers employed belonged to
different classes and therefore caution should be followed when their properties are
compared. MacDiarmid’s polymer [161] in the form of compressed pellet was in the EB-I
class, whereas the thin film cast from NMP utilized in the current project belonged to the
EB-II class. Jozefowicz et al. [234] showed that for the emeraldine base in the class II,
the onset of conductivity is even more difficult to accomplish and occurs at lower pHs. A
comparison of the two responses reported by MacDiarmid et al. [161] and Jozefowicz et
al. [234] is presented on the same graph in Figure 37.
Jozefowicz et al. [234] published just the “percent doping vs. pH” graph but also
considered that the protonation first occurs in the amorphous part of the EB-II crystal
structure, at pHs below 4, (i.e. a fact already suggested by MacDiarmid and Epstein [182]
and supported by the already published X-ray data [206,222]). Moreover, Jozefowicz et
al. [234] showed that the crystalline part of the EB-II transforms to the respective ES-II
crystal structure only at pHs equal to or lower than 2.3, or respectively when the doping
level in the polymer reaches 25%, in accordance to an early finding of MacDiarmid et al.
[182].
113
00.511.522.533.544.555.56
pH
Conductivity (S/cm)
MacDiarmid's work (compressed
pellet; EB-I class)
10-11
10-10
10-9
10-7
10-8
10-6
10-5
10-4
10-3
10-2
10-1
100
101
This work (NMP plasticized thin
film; EB-II class)
Figure 36. Comparison of the conductivity responses as a function of pH for the two classes of emeraldine base: the EB-I reported by MacDiarmid [161] and the EB-II
employed in the current project.
Therefore, it is not surprising that Figure 36 shows that the emeraldine base thin
film employed in the current project displayed an increase in conductivity at a pH level
lower than the one published by MacDiarmid et al. [161]. The second semicircle (i.e.
already mentioned in discussing Figure 34) starts forming at a pH level of HCl in water
of 5.03 for the sensor employed in these measurements and grows as the pH level of the
protonating solution is progressively lowered.
The appearance of the second semicircle is displayed in Figure 38. A zoom into
the low frequency area is displayed in Figure 39, showing that the low frequency random
noise align into a horizontal line at pH = 5.03, which gradually transforms into a
semicircle as the pH level is further decreased.
114
-101234567pH
% Doping
0
10
20
40
30
50
Jozefowicz's work (compressed
pellet, EB-II class)
MacDiarmid's work (compressed
pellet, EB-I class)
Figure 37. Comparison of the doping percentage vs. pH responses for the two classes of emeraldine base reported in the literature: MacDiarmid [161] and Jozefowicz [234]; The data points shown on the graph belong to MacDiarmid, being reproduced by DataThief
software close to the one originally published.
In all this time, the total resistance of the sensor given by its intercept with the
horizontal axis remains unchanged until the pH of the protonating bath reaches 2.25. A
measurement of the resistance performed with a DC multimeter would indicate that the
onset of conductivity increase does not occur until the pH of the solution reaches 2.25,
since the sensor would have displayed the same total resistance over entire pH range
employed. In fact, MacDiarmid et al. [160-162], Epstein and MacDiarmid [233] and
Jozefowicz et al. [234], all measured the conductivity by the four-probe technique, which
is a direct-current method of measurement.
115
Figure 38. The formation and evolution of the second semicircle with variation of pH
level.
-1800
-1600
-1400
-1200
-1000
-800
-600
-400
-200
0
14000 14500 15000 15500 16000 16500
Z' (ohm)
Z" (ohm)
pH=4.22
pH=4.52
pH=4.78
pH=5.03
pH=5.48
Figure 39. Appearance of the second semicircle as a horizontal line at pH=5.03 and its
subsequent evolution with decreasing pH level.
-8000
-6000
-4000
-2000
0 0 2000 4000 6000 8000 10000 12000 14000 16000
Z' (ohm)
Z" (ohm)
pH=4.22 pH=4.52
pH=4.78
pH=5.03
pH=3.88 pH=3.22
pH=2.25 pH=2.78
116
The onset of increase in conductivity (decrease in resistance) of the emeraldine
film at pH = 2.25 is displayed in Figure 40. At a pH level greater than 2.25, the total
resistance of the film was essentially constant. Once the pH of the protonating bath
reached 2.25, the resistance of the film started to decrease slowly. The slope in each pH
domain is different, meaning that the protonation became accelerated as the pH
progressively dropped, a fact expected since the amount of protons was increased by
lowering the pH.
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
0 100 200 300 400 500 600Time (min)
Resistance (ohm)
pH=1.75 pH=1.11pH=2.25pH=2.78
Figure 40. Resistance variation of the NMP-plasticized emeraldine base thin film with
time at different pH levels, showing the onset of conductivity change at pH = 2.25.
117
A graph showing both pH and the resistance of the first semicircle as a function of
hydrochloric acid concentration of protonating bath is presented in Figure 41. A
correlation between the measurements performed with two different techniques
(resistance measurements with interdigitated electrode and pH measurements with pH
meter) is observed. The fact that the resistance response of the emeraldine film follows
the titration curve of hydrochloric acid is an indication of the pH response of the polymer
that extends beyond the pH level of 4.
1.10
1.60
2.10
2.60
3.10
3.60
4.10
4.60
5.10
0.000 0.001 0.010 0.100 1.000
% HCl
pH
40
2040
4040
6040
8040
10040
12040
14040
Rhigh frequency (ohm)
Measured with pH meter
Measured with interdigitated electrode
Figure 41. Titration curve of hydrochloric acid in water and resistance response (Rhigh
frequency) of the emeraldine film as a function of hydrochloric acid concentration.
A plot of the capacitance values of the two semicircles given by the equivalent
circuit fit, as a function of pH is displayed in Figure 42. It is interesting to note that the
response can be separated in time dependent and time independent domains and the
118
border between the two is the pH of 2.25, the pH at which the onset in conductivity
increase occurs. The conduction mechanism which forms the initial semicircle (that starts
diminishing in magnitude as the pH is lowered below the value of 5) has associated with
it a virtually constant capacitance in the pH interval in which the overall conductivity of
the film is unchanged (i.e. pH between 2.25 and 5). The semicircle that starts forming at a
pH lower than 5 displays a progressively lower capacitance in this pH interval. When the
onset of conductivity change establishes, the capacitances associated with both
mechanisms increase with time of exposure in the respective pH.
0 1 2 3 4 5 6
pH
Capacitance (F)
Time dependent Time independent
10-8
10-7
10-6
10-5
10-4
10-9
Increasing exposure time
Chigh frequency
Clow frequency
Figure 42. Capacitance as a function of pH associated with the sensor response in hydrochloric acid.
119
Graphs showing the behavior of capacitances as a function of time spent in pH =
1.75 and pH = 1.11 displayed in Figure 43 and Figure 44 show parabolic dependence of
the capacitance as a function of square root of time, suggesting a diffusion controlled
process in these pH ranges.
0.0E+00
5.0E-09
1.0E-08
1.5E-08
2.0E-08
2.5E-08
3.0E-08
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00
Time1/2 (min
1/2)
Chigh freq. (F)
0.E+00
2.E-07
4.E-07
6.E-07
8.E-07
1.E-06
1.E-06
1.E-06
Clow frequency (F)
Chigh frequency
Clow frequency
Figure 43. Capacitance vs. square root of time for the emeraldine base thin film equilibrated for various times in pH = 1.75
120
0.0E+00
1.0E-06
2.0E-06
3.0E-06
4.0E-06
5.0E-06
0 1 2 3 4 5 6 7 8 9
Time 1/2
Capacitance (F)
Chigh frequency
Clow frequency
Figure 44: Capacitance vs. square root of time for the emeraldine base thin film
equilibrated for various times in pH = 1.11 The time constant (slope of the lines) differs in the two pHs, indicating that the diffusion
process (of chlorine ions) is pH dependent, as expected. Figure 43a in appendix display
the time dependence of capacitances of the two conduction mechanisms in pH levels of
1.75 and 1.11 on the same graph. Figure 43 and Figure 44 show that even in the same pH
level the time constant changes, suggestive to the fact that the achievement of equilibrium
protonation in each pH is dependent on diffusion of counter ions near the charged
portions of the emeraldine chain to maintain chain neutrality.
Impedance spectroscopy, due to the size and cost of equipment, is not amenable
for use in a practical sensor. However, the impedance measurement at a fixed frequency
would only require a simple oscillating circuit, which may be feasible for some
121
applications. The impedance of emeraldine base film at a specific frequency is defined
[209] as:
22 "' ZZZ += (30)
The impedance of two-different emeraldine base films measured at two different fixed
frequencies (1 Hz and 3.2 kHz) in different pHs in hydrochloric acid is displayed in
Figure 45.
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
2.533.544.555.5
pH
Impedance (ohm)
Impedance at 1 Hz
Impedance at 3.2 kHz
Figure 45. Impedance as a function of pH of hydrochloric acid solution for two fixed frequencies.
A graph showing the conductivity response of another emeraldine base thin film
prepared from the same stock solution is inserted in appendix as Figure 43b. The later
sensor having characteristic a conductivity of 4.78×10-3 S/cm (i.e. calculated with
122
equation (28)) right after equilibration in pure deionized water, displayed the onset of
increase in conductivity at a pH of hydrochloric acid of 2.75, and started displaying the
second semicircle at a higher pH (i.e. 5.85) than the sensor previously described. An
explanation of this discrepancy could rest in local arrangements difference, (i.e. finely
distributed, less scattered metallic islands in the insulating matrix, according to the
morphological model of emeraldine base described before).
3.2.1.2. Conductivity Response in Carbonic Acid Solutions
The conductivity response of the sensor with time and pH level in various
measurement conditions is presented in Figure 46 through Figure 48. Fitting the
measured data is an ambiguous step since there is sometimes more than a single
equivalent circuit that can fit the experimental data [209]. From the multitude of
equivalent circuits tested, the sequence of R-C described in Figure 16 and Figure 18
depicted the measured data on both criteria of best visual fit and low fitting error. Figure
46 displays the sensor resistance variation as a function of time in the various testing
conditions depicted on graph. The bubbling of 100% carbon dioxide in pure water lowers
the pH to a level of 3.95-4.0.
The removal of the aqueous carbon dioxide can be accomplished by bubbling an
inert gas through the solution, leading to an increase in pH displayed by carbonic acid
solution with time. The argon gas coming from a tank displaying by manufacturer a
purity of 99.9% lowered the pH of pure water to 4.65, indicating that some form of acidic
impurity was present. The purification of the gas in this particular cylinder was
accomplished by passing the gas through a 1 molar aqueous NaOH solution, an event that
123
led to the expected increase in the pH of aqueous solution. These two cycles are
displayed on Figure 48 as “Ar-unpurified” and “Ar-purified” respectively.
0
5000
10000
15000
20000
25000
0 50 100 150 200 250 300 350 400 450 500
Time (min)
Resistance (ohm)
H2OCO
2
H2O
CO
2 stopped
Ar purified
CO2 H2O CO2 CO
2
H2O
Ar unpurified
Rtotal equiv. circuit
Rhigh frequency equiv. circuit
Rlow frequency equiv. circuit
Figure 46. Sensor resistance in alternating cycles involving alternating bubbling of CO2, purified and unpurified Ar through water. Graph labels: “H2O”- pure water exposed to
room atmosphere air, “CO2”, “Ar purified” and “Ar-unpurified”- bubbling of respective gas in pure water. The equivalent circuit model gives the data points on graph.
The bubbling of unpurified argon gas had the same effect of bubbling carbon dioxide, i.e.
lowering the pH of solution and increase in the resistance of the low frequency
semicircle, suggestive to the fact that the sensor does not detect carbon dioxide but is
actually a pH based sensor.
On the graph in Figure 47 are displayed the resistances of the two semicircles
obtained as a geometrical intercept with horizontal axis, as well as the total resistance
obtained by adding these two values. At any pH greater than 5 the impedance spectra
consist of a single semicircle, therefore the total resistances given by means of
124
geometrical and equivalent circuit considerations are equal. Moreover, the low frequency
resistance is zero and high frequency resistance equals the total resistance. At pH values
lower than 5, the total resistance given by the geometrical intercept is greater than the
total resistance given by the equivalent circuit, due in fact to overlapping of the two
semicircles. However the difference between Rlow frequency and Rhigh frequency set of values
obtained in equivalent circuit and geometrical considerations is systematic.
Figure 48 depicts the response of the sensor in the above mentioned conditions as
a function of pH of aqueous solution. In addition, Figure 48 shows that the difference
between the high and the low frequency semicircles is greater when the pH of the
solution employed is higher. When the pH of the solution progressively decreases, the
low frequency semicircle increases on the expense of the high frequency semicircle and
the difference in resistance between the two diminishes.
The response time of the sensor to changes in pH level could not be accurately
determined by impedance spectroscopy measurements since an impedance measurement
takes about 3 minutes to complete, for a frequency sweep between 3.2 × 107 and 1 Hz.
125
0
5000
10000
15000
20000
25000
0 50 100 150 200 250 300 350 400 450 500
Time (min)
Resistance (ohm)
H2OCO
2
H2O
CO
2 stopped
Ar purified
CO2 H2O CO2 CO
2
H2O
Ar unpurified
Rtotal geometrical
Rhigh frequency geometrical
Rlow frequency geometrical
Figure 47. Sensor resistance in alternating cycles involving alternating bubbling of CO2, purified and unpurified Ar through water. Graph labels: “H2O”- pure water exposed to
room atmosphere air, “CO2”, “Ar purified” and “Ar-unpurified”- bubbling of respective gas in pure water. The geometrical intercept gives the data points on graph.
126
0
5000
10000
15000
20000
25000
3.753.954.154.354.554.754.955.155.355.555.75
pH
Resistance (ohm)
equiv. circuit
geometrical
geometrical
low frequency
high frequencyRtotal equiv. circuit
Rhigh frequency equiv. circuit
Rlow frequency equiv. circuit
Rhigh frequency geometrical
Rlow frequency geometrical
Rtotal geometrical
Figure 48. Sensor resistance as a function of pH, in alternating cycles involving alternating bubbling of CO2, purified and unpurified Ar through water.
3.2.2. Complexation Effect Study
The carbon dioxide detection mechanism is not based on the overall conductivity
change of the film. This is expected since the emeraldine base cannot be protonated at
pHs greater than 4. Nevertheless, even if no protonation of imine sites is expected from
carbonic acid, the carbonium ion (HCO3-) being a complexation agent could induce
complexation to the imine sites. In fact, any protonation or complexation of the chain
would produce the following events [213,215]:
1) an increase in the relative intensity of the peak centered at 1165 cm-1 (i.e. due
to the N=Q=N stretching mode) with respect to the peak centered at 1503 cm-1
(i.e. attributed to C=C stretching mode in benzenoid ring);
127
2) a progressive shift of the 1165 cm-1 peak to lower wavelength;
3) an increase in intensity of the ~1308 cm-1 peak (i.e. attributed to C-N type
stretching vibration).
Regarding the N=Q=N stretching mode, the fully protonated emeraldine salt has this peak
centered at 1144 cm-1 but intermediate levels of protonation display this peak at
intermediate wavelength between 1165 and 1144 cm-1, whereas the unprotonated
emeraldine base displays this peak centered at ~1165 cm-1
The characteristic FTIR spectra of the dried film is presented below in Figure 49.
The peak assignment displayed on graph is in excellent agreement with values reported
by Chen et al. [215] for emeraldine base films. Figure 50 presents the comparison of the
FTIR spectra between the dried film equilibrated in water and the same film immersed
for 10 minutes in H2CO3 solution of pH = 3.95.
128
0
10
20
30
40
50
60
70
80
40090014001900240029003400
Wavelength (cm-1)
Transm
ittance (a.u.)
3380 cm-1: Symmetrical N-H
Stretching in B-NH-B mode +
Symmetrical N-H Stretching
3296 cm-1: Hydrogen
bonded N-H
Stretching
3170 cm-1: N-H Stretching
in Q=NH mode
830 cm-1: C-H out-of-
plane bending on 1,4-
ring
1503cm-1: Stretching
in C=C mode-Benzenoid
1593cm-1: Stretching
in C=C mode-Quinoid
2930, 2868 cm-1:
Impurity or sum freq.
1380 cm-1: C-N
stretching in QBtQ
1308 cm-1: C-N stretching in
QBcQ + QBB + BBQ
1165 cm-1: N=Q=N
stretching mode
1115, 1060,
970 cm-1:
C-H in-
plane
bending on
1,2,4-ring
739, 690 cm-1: C-H
out-of-plane bending
on 1,2-ring
1657cm-1: C=O
Stretching
in NMP
Figure 49. FTIR spectrum of the dried emeraldine base thin film.
129
0
10
20
30
40
50
60
70
80
40090014001900240029003400
Wavelength (cm-1)
Transm
ittance (a.u.)
EB in pure water exposed to air
EB-10 min in carbonic acid
1594 cm-1: C=C
Stretching-Quinoid1502 cm
-1: C=C
Stretching-Benzenoid
1164 cm-1: N=Q=N
Stretching
1660 cm-1: C=O
Stretching in NMP
Figure 50. Comparison of the FTIR spectra between the dried film and the film immersed for 10 minutes in carbonic acid solution of pH = 3.95.
None of the above-described effects was observed in this work; therefore, no
protonation or complexation of the imine sites could be detected by FTIR and attributed
to the exposure of emeraldine base film to carbonic acid. Thus, the FTIR complements
the impedance measurements, showing that no protonation and/or complexation of the
emeraldine chain is possible with carbonic acid.
However, Chen et al. [215] observed that protonation in emeraldine base thin
films is more difficult to accomplish than in the respective powder pellets, due to the
plasticizing effect of NMP residue. In the next development it is necessary to assess
whether the lack of protonation and/or complexation of imine sites is due to the inherent
NMP content in the film (i.e. the residual NMP is expected to be somewhere between 10
130
and 18 %, according to numerous reports in the literature [145,160-162,167,212,215,
216]). Therefore an acid base treatment was performed on emeraldine base film in order
to remove the residual NMP (i.e. displayed on the graph in Figure 49 by the ~1657 cm-1
C=O stretching vibration). The FTIR spectra were recorded at various pH levels during
the acid exposure, to assess also the onset of conductivity in emeraldine film.
A comparison of the spectra taken at two pH levels in aqueous hydrochloric acid
(i.e. 3.88 and 1.75) displayed in Figure 51 show that the two spectra are identical,
therefore no protonation had begun at the pH of 1.75.
0
10
20
30
40
50
60
70
40090014001900240029003400
Wavelength (cm-1)
Transm
ittance (a.u.)
pH=3.88
pH=1.75
Figure 51. Comparison of the FTIR spectra for emeraldine base treated in HCl solution of pH = 3.88 and 1.75.
On the other hand, the exposure of emeraldine base to the next employed pH of 1.11
displayed in Figure 52 shows that the protonation in emeraldine base had started at a pH
131
intermediate between 1.75 and 1.11. This protonation event is indicated by the growing
of both peaks centered at 1302 cm-1 and 1165 cm-1 and the shift of the 1165 cm-1 peak to
1142 cm-1, as displayed in Figure 52. The peak at 1142 cm-1 becomes the highest peak in
the spectrum, a fact already reported elsewhere [149,160,162,167,239]. Due to its dense
morphology and a greater thickness compared to the film deposited on interdigitated
electrode (which showed an onset of conductivity at pH = 2.25), no conductivity change
was detected on FTIR spectra at pHs greater than 1.75 for this film deposited on glass
slide, even though the films were deposited from the same stock solution.
0
1
2
3
4
5
6
45095014501950245029503450
Wavelength (cm-1)
Transm
ittance (a.u.)
1570cm-1: Stretching
in C=C mode-Quinoid
1482cm-1: Stretching
in C=C mode-Benzenoid 1302 cm-1: C-N stretching in
QBcQ + QBB + BBQ
810 cm-1: C-H out-of-
plane bending on 1,4-
ring
1142 cm-1: N=Q=N
stretching mode
Figure 52. FTIR spectrum of the EB film treated in pH = 1.11. A comparison of the FTIR spectra taken with the film equilibrated in pure water
after the acid base treatment and for 10 minutes in carbonic acid solution of pH = 3.95 is
132
displayed in Figure 53. Inspecting the graph can be concluded that there is no indication
that the complexation of imine sites occurred. The only accomplishment of the acid-base
treatment is represented by the removal of the NMP residue, proved by the disappearance
of the peak centered at ~1660 cm-1 and the persistence of a small shoulder centered at
~1665 cm-1 instead.
0
5
10
15
20
25
30
40090014001900240029003400
Wavelength (cm-1)
Transm
ittance (a.u.)
EB in pure water exposed to air
EB-10 min carbonic acid
1665 cm-1: C=O
stretching in NMP
Figure 53. Comparison of the FTIR spectra between the films equilibrated in water and in carbonic acid solution of pH = 3.95 for 10 minutes. The spectra is recorded after the acid-
base treatment of the film was accomplished. 3.2.3. Response of the Sensor in Ar-CO2 mixtures
Having the sensor response assessed in hydrochloric solutions of various pHs, in
aqueous solutions of saturated carbonic acid (pH ≅ 4), and in solutions of intermediate
pH generated through bubbling of unpurified argon (pH ≅ 4.65), attempts were made to
133
assess the minimal concentration of carbon dioxide that the sensor could detect.
Therefore, various mixtures of ultra pure carbon dioxide gas and purified argon gas were
bubbled through pure deionized water and the respective pH generated measured.
The sensor employed for these measurements was the sensor whose conductivity
response to hydrochloric acid is displayed in Figure 43b in appendix. Between the
measurements in hydrochloric acid and in carbonic acid, 11 months interval elapsed. All
this time the sensor was preserved in water in a sealed beaker to avoid evaporation. After
11 months of preservation, the conductivity displayed by the sensor in pure water of pH =
7 in which was equilibrated increased by a factor of 5.
The sensor tested in carbonic acid solutions of various pHs displayed a second
semicircle formation at a pH of 5.85, generated by bubbling a mixture of Ar-0.015% CO2
for 30 minutes in deionized water. The pH at which this sensor formed the second
semicircle is in fact 0.83 pH units higher than the sensor described throughout the
“immerse sensor part” of the current project. Figure 54 presents the resistance response of
the sensor with various pH levels generated by bubbling the two-gases mixture in pure
water.
134
0
100
200
300
400
500
3.754.004.254.504.755.005.255.505.756.006.25
pH
Resistance (ohm)
Rtotal
Rhigh frequency
Rlow frequency
Figure 54. Resistance of emeraldine base thin film as a function of pH of carbonic acid, generated through bubbling of various mixtures of carbon dioxide and argon in
deionized water.
The appearance of the second semicircle at 0.0150% CO2 (i.e. 150 ppm CO2) is displayed
in Figure 55.
135
-200
-180
-160
-140
-120
-100
-80
-60
-40
-20
0
0 50 100 150 200 250 300 350 400 450
Z' (ohm)
Z" (ohm)
-5
-4
-3
-2
-1
0425 428 431 434Z' (ohm)
Z" (ohm)
pH =
pH = 5.75
pH
pH = 5.5
pH = 6.0
pH = 5.85
pH = 5.5
pH = 4.10
pH = 6.00
Figure 55. The appearance of the second semicircle at a pH of 5.85, generated by bubbling a mixture of Ar-0.015% CO2 for 30 minutes in pure deionized water
Figure 56 displays the evolution of the resistance of the initial semicircle
(denominated as Rhigh frequency, measured with interdigitated electrode) with concentration
of carbon dioxide in the gas mixture displayed on logarithmic scale. On the same graph is
plotted the pH generated in the aqueous solution (measured with pH meter) as a function
of CO2 concentration. The measurements with these two techniques were performed
simultaneously. The good correlation between these results is suggestive to a pH
response of the emeraldine base thin film, a fact already observed in hydrochloric acid
solutions (Figure 41).
136
3.95
4.45
4.95
5.45
5.95
% CO2
R high frequency (ohm)
389
394
399
404
409
414
419
424
429
434
pH
Rhigh frequency
pH
0.01 0.1 1 10 1000.001
Figure 56. The high frequency resistance of the emeraldine thin film as a function of carbon dioxide concentration in Ar-CO2 mixture (black) and pH response of solution to
bubbling carbon dioxide (red)
The impedance of the emeraldine base film measured at two different fixed
frequencies (1 Hz and 3.2 kHz) in carbonic acid (Ar-CO2 bubbled through water) is
shown in Figure 57. The two sets of measurements (in carbonic acid, Figure 57 and
hydrochloric acid, Figure 45) were performed with two different interdigitated electrodes
so the magnitudes of the resistances were different. The fixed frequency used to calculate
the impedance was chosen to maximize the dynamic range of the sensor. In both cases
the high frequency measurement produces a response not observable at 1 Hz, indicating
that an oscillating circuit could be used to sense changes not detectable with DC
measurements.
137
The appearance of the second semicircle may be consistent with the
morphological model of emeraldine base reported in the literature [145,182,194,207].
The display of two semicircles on the impedance spectra would therefore be
representative of protonation in the insulating matrix, and the resistance decrease of the
original semicircle with lowering the pH (e.g. as displayed in Figure 38) would suggest
the progressive increase in protonation in this initially insulating region. The overall
conductivity of the film increases only when the level of protonation in this insulating
region reaches a value that allows three-dimensional delocalization of electrons and
electron hopping between metallic islands.
370
380
390
400
410
420
430
440
450
%CO2
Impedance (ohm)
0.010.001 0.1 1 10 100
Impedance at 1 Hz
Impedance at 3.2 kHz
Figure 57. Impedance of emeraldine base thin film for two fixed frequencies as a function of carbon dioxide concentration in argon-CO2 mixture; a correlation between the carbon
dioxide concentration and pH generated in pure water is displayed in Figure 56.
138
At pHs greater than the level necessary for the onset of conductivity change, the
film response is governed by the electron transfer mechanism (the virtually constant
capacitance of the first semicircle in these pHs displayed in Figure 42 being suggestive to
no limitation in the counter ion diffusion). When the onset of conductivity increase is
established, the conduction mechanism is governed by both electron transfer (the shape
of impedance curve is a semicircle, suggestive to an electron transfer as the slowest
mechanism [209]) and diffusion (the first semicircle capacitance has a parabolic
dependence, suggestive to a diffusion controlled process).
The formation of the second semicircle appears to depend on the initial
conductivity of the emeraldine base film, since a film displaying an initial conductivity of
4.78×10-3 S/cm formed the second semicircle at a pH of 5.85.
3.2.4. Conclusions
The emeraldine base sensor developed in the second part of the project presents
some interesting features, not yet reported in the literature. First, the appearance of the
second semicircle in the impedance spectra, due to an activated conduction mechanism at
pHs lower than 6.0 was not reported yet. The overall resistance of the film follows
closely the curve depicted by Chiang, Huang and MacDiarmid [160-162]. The use of
impedance spectroscopy makes the emeraldine base thin film act as a pH sensor even at
pH levels where there is virtually no protonation and therefore no measurable increase in
the film conductivity.
139
The emeraldine base can be used for detection of carbon dioxide in water
equilibrium, but the response is governed by the pH change induced by the formation of
carbonic acid.
The sensor needs water abundantly, which is insufficiently provided by the matrix
of polyvinyl alcohol. Depending on the local arrangement in the emeraldine film, the
onset of second semicircle appearance (i.e. the onset of protonation in the insulating
region of the polymer) can occur at various pHs, a value as high as 5.85 being recorded
with this type of sensor.
The sensor developed in the second part of this work can find its suitability in
monitoring the pH change of liquid samples. The appearance of the second semicircle
would be indicative of pH dropping below a specific threshold.
Single frequency AC impedance measurements allows for the use of emeraldine
base polyaniline (and probably other conductive polymers) in pH sensing applications at
protonation levels above the threshold for generation of a total conductivity change and
thus expands opportunities for sensor development.
140
4. FUTURE WORK
Future research should be aimed towards analyzing the conductivity properties of
emeraldine base prepared in different conditions given by powder fabrication history,
molecular weight, NMP solution characteristics (water uptake, concentration), etc.
The current project cannot conclude if there is a direct correlation between the
initial conductivity of the emeraldine film and the pH level at which the second
semicircle starts forming, therefore future research should be accomplished in order to
elucidate this behavior of emeraldine base. In addition, the extent of increase in
conductivity during preservation in water should be evaluated for emeraldine films
produced from stock powders with different fabrication history.
The effect of low pressure over the conductivity of emeraldine base should be
carefully investigated since cycling the polymer between low pressure and atmospheric
pressure could induce a conductivity variation and therefore preclude the use of vacuum
pump as an experimental technique.
The usefulness of poly(vinyl alcohol) as a water pool should be further evaluated
in order to assess whether the removal of NMP residue play a part in water uptake
characteristics of PVA. This will allow the emeraldine base thin film to detect carbon
dioxide without the need of inserting the sensor in water.
141
The fabrication of emeraldine base with 50% oxidation states should be
attempted, by controlling the oxygen partial pressure during the polymerization reaction.
Then the conduction properties of this material should be carefully investigated.
142
5. REFERENCES
[1] Meridian Diagnostics EMISA Test Specifications Sheet, E Coli O157:H7 (1988). [2] Neogen Corporation, “Reveal for Salmonella Sensitivity Tests” Test Specifications Sheet, Salmonella (1998). [3] J.P.E. Anderson, K.H. Domsch, “Measurement of Bacterial and Fungal Contributions to Respiration of Selected Agricultural and Forest Soils”, Canadian J. Microbiol., 21 (1975) 314-321. [4] J.P.E. Anderson, K.H. Domsch, “A Physiological Method for the Quantitative Measurement of Microbial Biomass in Soils”, Soil Biol. Biochem., 10(3) (1978) 215-221. [5] T. Harden, R.G. Joergensen, B. Meyer, V. Wolters, “Soil Microbial Biomass Estimated by Fumigation-Extraction and Substrate-Induced Respiration in Two Pesticide-Treated Soils”, Soil Biol. Biochem. 25(6) (1993) 679-683. [6] O. Dilly, J.C. Munch, “Microbial Biomass Content, Basal Respiration and Enzyme Activities During the Course of Decomposition of Leaf Litter in a Black Alser (Alnus Glutinosa L. Gaerten) Forest”, Soil Biol. Biochem., 28(8) (1996) 1073-1081. [7] T. Nordgren, E. Baath, B. Sonderstrom, “Evaluation of Soil Respiration Characteristics to Assess Heavy Metal Effects on Soil Microorganisms Using Glutamic Acid as a Substrate”, Soil Biol. Biochem., 20(6) (1988) 949-954. [8] K. Chander, P.C. Brookest, “Microbial Biomass Dynamics During the Decomposition of Glucose and Maize in Metal-Contaminated and Non- Contaminated Soils”, Soil Biol. Biochem. 23(10) (1991) 917-925. [9] P.A. Hartman, B. Swaminathan, M.S. Curiale, R. Firstanberg-Eden, A.N. Sharpe, N.A. Cox, D.Y.C. Fung, C. Goldschmidt, “Rapid Methods and Automation”, in C. Vanderzant and D.F. Splittstosser (Eds.), Compendium of Methods for the Microbiological Examination of Foods (Am. Publ. Health Assoc., Arlington, VA, 1992), 665-746. [10] C.H. Threlkeld, “Detection of Microbial Contamination Utilizing an Infrared CO2 Analyzer”, J. Food Sci., 47 (1982) 1222-1225.
143
[11] D.E. Townsend, A. Naqui, “Comparison of SimPlate Total Plate Count Test with Plate Count Agar Method for Detection and Quantitation of Bacteria in Food”, J. AOAC Int., 81(3) (1998) 563-569. [12] C.F. Smith, D.E. Townsend, “A New Medium for Determining the Total Plate Count in Food”, J. Food Protection, 62(12) (1999) 1404–1410. [13] S.Y. Chew, Y.H.P. Hsieh, “A Rapid CO2 Evolution Method for Determining Shelf Life of Refrigerated Catfish“, J. Food Sci., 63(5) (1998) 768-771. [14] Y.P. Hsieh, Y.H.P. Hsieh, “Simple System for a Rapid Determination of Carbon Dioxide Evolution Rates”, J. AOAC Internat., 81(3) (1998), 652-656. [15] Y.P. Hsieh, Y.H.P. Hsieh, “Determination of Carbon Dioxide Evolution Rates Using a Novel Noninstrumental Microrespirometer”, J. AOAC Internat., 83(2) (2000) 277-281. [16] H.H. Mobius, “Galvanic Solid Electrolyte Cells for the Measurement of CO2 Concentrations”, J. Solid. State. Electrochem., 8 (2004) 94-109. [17] N. Yamazoe, N. Miura, “Gas Sensors Using Solid Electroltes”, MRS Bulletin, June (1999), 37-43. [18] R.W. Stow, R.F. Baer, B. Randall, “Rapid Measurement of the Tension of Carbon Dioxide in Blood”, Arch. Phys. Med. Rehabil., 38 (1957) 646-650. [19] W. Severinghaus, A.F. Bradley, “Electrodes for Blood pO2 and pCO2 Determination”, J. Appl. Physiol., 13 (1958) 515-520. [20] C.E.W. Hahn, “Electrochemical Analysis of Clinical Blood-Gases, Gases and Vapors”, Analyst, 123(6) (1998), 57R-86R. [21] C.R. Caflisch, N.W. Carter, “A Micro pCO2 Electrode”, Anal. Biochem., 60 (1974) 252–257. [22] A.K. Covington, R.A. Robinson, M. Sarbar, “Determination of Carbonate in Presence of Hydroxide; Part I: Analysis of First Derivative Potentiometric Curves”, Anal. Chim. Acta, 100 (1978) 367-374. [23] A.K. Covington, R.A. Robinson, M. Sarbar, “Determination of Carbonate in Presence of Hydroxide; Part II: Evidence for the Existence of Novel Species from First Derivative Potentiometric Titration Curves”, Anal. Chim. Acta, 130 (1981) 93-102.
144
[24] A.K. Covington, R.N. Goldberg, M. Sarbar, “Computer Simulation of Titration Curves with Application to Aqueous Carbonate Solutions” Anal. Chim. Acta, 130 (1981) 103-109. [25] D. de Beer, A. Glud, E. Epig, M. Kuhl, “A Fast Responding CO2 Microelectrode for Profiling Sediments, Microbial Mats and Biofilms ”, Limnol. Oceanogr., 42(7) (1997) 1590-1600. [26] S. Hanstein, D. de Beer, H.H. Feele, “Miniaturized Carbon Dioxide Sensor Designed for Measurements Within Plant Leaves”, Sens. Actuators B, 81 (2001) 107-114. [27] S. Yao, M. Wang “Electrochemical Sensor for Dissolved Carbon Dioxide Measurement”, J. Electrochem. Soc., 149 (1) (2002) H28-H32. [28] H. Beyenal, C.C. Davis, Z. Lewandowski, “An Improved Severinghaus Type Carbon-Dioxide Microelectrode for Use in Biofilms”, Sens. Actuators B 97 (2004) 202-210. [29] T. Sekiguchi, Y. Nagai, T. Makino, K. Ohno, M. Nakamura, H. Hosaka, H. Sakio, S. Ohtsu, H. Takahashi, “Gastric PCO2 Monitoring System Based on a Double Membrane Type PCO2 Sensor”, Sens. Actuators B, 49 (1998), 171-178. [30] H. Suzuki, H. Arakawa, S. Sasaki, I. Karube, “Micromachined Severinghaus- Type Carbon Dioxide Electrode”, Anal. Chem., 71 (1999) 1737–1743. [31] W.J. Cai, C.E. Reimers “The Development of pH and pCO2 Micoelectrodes for Studying the Carbonate of Pore Waters Near the Sediment-Water Interface”, Limnol. Oceanogr., 38 (1993) 1762-1773. [32] Y.J. Zheng, K.M. Merz, Mechanism of the Human Carbonic Anhydrase II Catalyzed Hydration of Carbon Dioxide, J. Am. Chem. Soc. 114 (1992) 10498– 10507. [33] C. Cammaroto, L. Diliberto, M. Ferralis, R Manca, A. Sanna, M. Giordano, “Use of Carbonic Anhydrase in Electrochemical Biosensors for Dissolved CO2
”, Sens. Actuators B, 48 (1998) 439–447. [34] P. Zhao, W.J. Cai, “An Improved Potentiometric pCO2 Microelectrode, Anal. Chem. 69 (1997) 5052–5058. [35] F.J.W. Roughton, A.M. Clark, “Carbonic Anhydrase”, in: J.B. Summer, K. Myrback (eds.), “The Enzymes. Chemistry and Mechanism of Action”: Part 2, Vol. 1, Academic Press, New York, 1951, pp. 1250-1265
145
[36] T. Komada, C.E. Reimers, S.E. Boehme, “Dissolved Inorganic Carbon Profiles and Fluxes Determined Using pH and pCO2 Microelectrodes”, Limnol. Oceanogr., 43(5) (2001) 769–781. [37] H.A. Himpler, S.F. Brand, M.J.D. Brand, “Conductimetric Gas Sensor for Carbon Dioxide” Anal. Chem., 50 (1978) 1623. [38] C.M. Santegoeds, A. Schramm, D. de Beer, “Microsensors as a Tool to Determine Chemical Microgradients and Bacterial Activity in Wastewater Biofilms and Flocs”, Biodegradation 9 (1998) 159–167. [39] A.N. Bezbaruah, T.C. Zhang, “Fabrication of Anodically Electrodeposited Iridium Oxide Film pH Microelectrodes For Microenvironment Studies”, Anal. Chem., 74 (2002) 5726–5733. [40] P. Van Houdt, Z. Lewandowski, B. Little, Iridium Oxide pH Microelectrode”, Biotechnol. Bioeng., 40 (1992) 601–608. [41] T. Ishiji, K. Takashashi, A. Kira, “Amperometric Carbon Dioxide Gas Sensor Based on Electrode Reduction of Platinum Oxide”, Anal Chem., 65 (1993) 2736- 2739. [42] W. J. Albery, P. Barron, “A Membrane Electrode for the Determination of CO2 and O2”, J. Electroanal. Chem., 138 (1982) 79-87. [43] F. Qian, J. Lu, Z. Zhou, C. Cha, “Combined Amperometric Sensors for Simultaneous Measurement of carbon Dioxide and Oxygen”, Sens. Actuators B, 17 (1993) 77-83 [44] H. Suzuki, T. Hirakawa, T. Hooshi, H. Toyooka, “Micromachined Sensing Module for pO2, pCO2 and pH and its Design Optimization for Practical Use”, Sens. Actuators B, 76 (2001), 565-572. [45] Z.B. Zhou, Q.H. Wu, C.C. Liu, “Microfabricated Thin-Film Microelectrode for Amperometric Determination of CO2 in the Gas Phase”, Sens. Actuators B, 21 (1994) 101-108. [46] J. Evans, D. Pletcher, P.R. Warburton, T.K. Gibbs, “Amperometric Sensor for Carbon Dioxide: Design, Characteristics and Performance”, Anal. Chem., 61 (1989) 577-580. [47] J. Evans, D. Pletcher and P.R.G. Warburton, T.K. Gibbs, “A New Electrochemical Sensor for Carbon Dioxide Part 2. Study of the Sensor Chemistry”, J. Electroanal. Chem., 262 (1989) 119-129.
146
[48] T. Ishiji, D.W. Chipman, T. Takahashi, K. Takahashi, “Amperometric Sensor for Measurement of Dissolved Carbon Dioxide in Seawater”, Sens. Actuators B, 76 (2001) 265-269. [49] R. Fasching, F. Keplinger, G. Hanreich, G. Hobst, G. Urban, F. Kohl, R. Chabicovsky, “A Novel Miniaturized Sensor for Carbon Dioxide Dissolved in Liquids”, Sens. Actuators B, 78 (2001) 291-297. [50] R. Fasching, F. Kohl, G. Urban, “A Miniaturized Amperometric CO2 Sensor Based on Dissociation of Copper Complexes”, Sens. Actuators B, 93 (2003) 197- 204. [51] J.D. Wadhawan, P.J. Welford, H.B. McPeak, C.E.W. Hahn, R.G. Compton “The Simultaneous Voltammetric Determination and Detection of Oxygen and Carbon Dioxide. A Study of The Kinetics of the Reaction Between Superoxide and Carbon Dioxide in Non-Aqueous Media Using Membrane-Free Gold Disc Microelectrodes”, Sens. Actuators B, 88 (2003) 40-52 [52] J.L. Roberts, D.T. Sawyer, “Voltammetric Determination of Carbon Dioxide Using Dimethylsulfoxide as a Solvent”, J. Electroanal. Chem., 9 (1965) 1–7. [53] L.V. Haynes, D.T. Sawyer, “Electrochemistry of Carbon Dioxide in Dimethyl Sulfoxide at Gold and Mercury Electrodes”, Anal. Chem., 39 (1967) 332. [54] T.E. Teeter, P. Van Rysselberghe, “Reduction of Carbon Dioxide on Mercury Cathodes” J. Chem. Phys., 22 (1954) 759-760. [55] J. Giner, “Electrochemical Reduction of CO2 on Platinum Electrodes in Acid Solutions”, Electrochim. Acta, 8(11) (1963), 857-865. [56] J. Giner, “The Anodic Oxidation of Methanol and Formic Acid and The Reductive. Adsorption of CO2 ”, Electrochim. Acta, 9(1) (1964), 63-77. [57] B.R. Eggins, J. McNeil, “Voltammetry of CO2. Part 1. General Survey of Voltammetry at Different Electrode Materials in Different Solvents”, J. Electroanal. Chem., 148 (1983) 17-24. [58] P.J. Welford, B.A. Brookes, J.D. Wadhawan, H.B. McPeak, C.E.W. Hahn, R.G. Compton, “The Electro-Reduction Of Carbon Dioxide in Dimethylsulphoxide at Gold Microdisc Electrodes: Current/Voltage Wave Shape Analysis”, J. Phys. Chem. B, 105 (2001) 5253–5261. [59] A. Gennaro, A.A. Isse, E. Viannello, “Solubility and Electrochemical Determination of CO2 in Some Dipolar Aprotic Solvents”, J. Electroanal. Chem., 289 (1990) 203–215.
147
[60] J.L. Roberts Jr., D.T. Sawyer, “Electrochemistry of Oxygen and Superoxide Ion in Dimethylsulfoxide at Platinum Mercury and Gold Electrodes”, J. Electroanal. Chem., 12 (1966) 90–101. [61] D.T. Sawyer, G. Chiericato Jr., C.T. Angelis, E.J. Nanni Jr., T. Tsuchiya, “Effects of Media and Electrode Materials on the Electroreduction of Dioxygen”, Anal. Chem., 54 (1982) 1720–1724. [62] J. Sobkowsky, A. Wieckowski, P. Zelenay A. Czerwinski, “Electrochemical Reduction of CO2 and Oxidation of Adsorbed Species on the Rhodium Electrodes”, J. Electroanal. Chem., 100 (1979) 781-790. [63] C.E.W. Hahn, “Techniques for Measuring the Partial Pressures of Gases in the Blood. Part I-In Vitro Measurements”, J. Phys. E, 13 (1980) 470-482. [64] C.E.W. Hahn “Techniques for Measuring the Partial Pressures of Gases in the Blood. Part II-In Vivo Measurements ”, J. Phys. E, 14 (1981) 783-797. [65] C.E.W. Hahn, H. McPeak, A.M. Bond, D. Clark, “The Development of New Microelectrode Gas Sensors: An Odyssey. Part 1. O2 and CO2 Reduction at Unshielded Gold Microdisc Electrodes”, J. Electroanal. Chem., 393 (1995) 61-68. [66] C.E.W. Hahn, H. McPeak, A.M. Bond, “The Development of New Microelectrode Gas Sensors: an Odyssey. Part 2. O2 and CO2 Reduction at Membrane-Covered Gold Microdisc Electrodes”, J. Electroanal. Chem., 393 (1995) 69-74. [67] H. McPeak, A.M. Bond, C.E.W. Hahn, “The Development of New Microelectrode Gas Sensors: an Odyssey Part IV. O2, CO2 and N2O Reduction at Unshielded Gold Microdisc Electrodes”, J. Electroanal. Chem., 487 (2000) 25-30. [68] W.J. Albery, D. Clark, H.J.J. Drummond, A.J.M. Coombs, W.K. Young, C.E.W. Hahn, “Pulsed Titration Sensors Part 1: A Breath-by-Breath CO2 Sensor ” J. Electroanal. Chem., 340 (1992) 99-110. [69] W.J. Albery, D. Clark, W.K. Young and C.E.W. Hahn, “Pulsed Titration Sensors Part 2. A General Theoretical Model for the Breath-by-Breath CO2 Sensor”, J. Electroanal. Chem., 340 (1992) 111-126. [70] J.D. Wadhawan, P.J. Welford, E. Maisonhaute, V. Climent, N.S. Lawrence, R.G. Compton, H.B. McPeak, C.E.W. Hahn, “Microelectrode Studies of the Reaction of Superoxide with Carbon Dioxide in Dimethyl Sulfoxide”, J. Phys. Chem. B, 105 (2001) 10659-10668.
148
[71] Z.B. Zhou, W.J. Liu, C.C. Liu, “Non-Invasive, Real-Time and Dynamic Monitoring of CO2 and O2 Simultaneously Using Modulated Potential Pulse- Amperometry: Coulometry”, Sens. Actuators B, 52 (1998) 219–225. [72] J.L. Roberts, T.S. Calderwood, D.T. Sawyer, “Nucleophilic Oxygenation of Carbon Dioxide by Superoxide Ion in Aprotic Media To Form the C2O6
2-
Species”, J. Am. Chem. Soc., 106 (1984) 4667-4670. [73] J.R. Cain, L.C. Maxwell, “An Electrolytic Resistance Method for Determining Carbon in Steel”, J. Ind. Eng. Chem., 11(9) (1919) 852-860. [74] I.B. Holm-Jensen, “A New Gas Absorption Device Its Application to Titrimetric and Conductometric Micro Determinations of Carbon Dioxide in Air”, Anal. Chim. Acta, 23 (1969) 13-27. [75] R. H. Maffly, “A Conductometric Method for Measuring Micromolar Quantities of Carbon Dioxide” Anal. Biochem., 23 (1968) 252-262. [76] D.B. James, “An Instrument for Gas Analysis”, Biomed. Eng., 4(3) (1969) 126- 133. [77] K. Lis, H. Acker, D.W. Lubbers, M.Halle, “A pCO2 Surface Electrode Working on the Principle of Electrical Conductivity”, Pflugers Arch., 381 (1979) 289-291. [78] S. Bruckenstein, J.S. Symanski, “Continuous Conductometric Sensor for Carbon Dioxide”, Anal. Chem., 58(8) (1986) 1766-1770. [79] A.R. Varlan, W. Sansen, “Micromachined Conductometric pCO2 Sensor”, Sensors and Actuators B, 44 (1997) 309-315. [80] S. Bruckenstein, J.S. Symanski, “Analytical Applications of Gas Membrane Electrodes”, J. Chem. Soc. Faraday Trans. 1, 82(4), (1986) 1105 – 1116. [81] Y. Saikai, Y. Sadaoka, M. Matsuguchi, H. Yokouchi, K. Tamai, “Effect of Carbon Dioxide on the Electrical Conductivity of Polyethylene Glycol-Alkali Carbonate Complex Film”, Mater. Chem. Phys., 42 (1995) 73-76. [82] Z.B. Zhou, W.J. Liu, C.C. Liu, “Studies on The Biomedical Sensor Techniques for Real-Time and Dynamic Monitoring of Respiratory Gases, CO2 and O2”, Sens. Actuators B, 65 (2000) 35–38. [83] Z.B. Zhou, L.D. Feng, W.J. Liu, Z.G. Wu, “New Approach for Developing Transient Electrochemical Multi-Component Gas Sensors”, Sens. Actuators B, 76 (2001) 605-609.
149
[84] B. van der Schoot, P. Bergveld, “The Coulometric Carbon Dioxide Sensor, a New Type of Chemical Transducer”, Proceedings of the Second International Meeting on Chemical Sensors, Bordeaux, 1986, 665–668. [85] W.J. Albery, M. Uttamlal, “A CO2 Titration Electrode Part I: Theoretical Description”, J. Appl. Electrochem., 24 (1994) 8-13. [86] W.J. Albery, M. Uttamlal, M.S. Appleton, N.J. Freeman, B.B. Kebbekus, M.D. Neville, “A CO2 Titration Electrode Part II: Development of the Sensor”, J. Appl. Electrochem., 24 (1994) 14-17. [87] T. Trapp, B. Boss, K. Cammann, E Schirmer, C. Berthold, “Development of a Coulometric CO2 Gas Sensor”, Sens. Actuators B, 50 (1998) 97-103. [88] K. Wiegran, T. Trapp, K. Cammann, “Development of a Dissolved Carbon Dioxide Sensor Based on Coulometric Titration” Sens. Actuators B, (1999) 120- 124. [89] H.C. Chiu, A.C.C. Tseung, “Electrochemical CO2 Sensor Based on CO2 Reduction and Subsequent CO Oxidation on Pt/WO3 Electrode”, J. Electrochem. Soc. Lett., 2(10) (1999) 540-541. [90] P. Bergveld, “Development, Operation and Application of the Field Effect Transistor as a Tool for Electrophysiology”, IEEE Trans. Biomed. Eng., BME 19 (1972) 342-351. [91] J. Shaw, “Bilayer Membrane for ISFETs to Eliminate CO2 Mediated pH Sensitivity”, Sens. Actuators B, 15-16 (1993) 81-85. [92] H. H. Van Den Vlekkert, N. F. De Rooij, A. Van Den Berg, A. Grisel, “Multi-Ion Sensing System Based on Glass-Encapsulated pH-ISFETs and a Pseudo-REFET”, Sens. Actuators B. 1 (1990) 395-400. [93] M. Trojanowicz, Z. Augustowska, W. Matuszewski, G.Moraczewska, A. Hulanicki, “A Potassium-Selective Electrode With Solid Internal Contact”, Talanta, 29 (1982) 113-117. [94] X. Li, E.M.I. Verpoorte, D. J. Harrison, “Elimination of Neutral Species Interference at the Ion-Sensitive Membrane/Semiconductor Device Interface”, Anal. Chem., 60 (1988) 493-498. [95] P. Herman, Z. Murtaza, J.R. Lakowicz, “Sensing of Carbon Dioxide by a Decrease in Photoinduced Electron Transfer Quenching”, Anal. Biochem., 272 (1999) 87–93.
150
[96] M.D. Marazuela, M.C. Moreno-Bondi, G. Orellana, “Enhanced Performance of a Fiber Optic Luminescence CO2 Sensor Using Carbonic Anhydrase”, Sens. Actuators B, 29 (1995) 126-131. [97] O.S. Wolfbeis, L.J. Weis, M.J.P. Leiner, W.E. Ziegler, “Fiber Optic Fluorosensor for Oxygen and Carbon Dioxide”, Anal. Chem., 60 (1988) 2028–2030. [98] E. M. Hampe, D. M. Rudkevich, “Exploring Reversible Reactions between CO2 and Amines”, Tetrahedron, 59 (2003) 9619-9625. [99] S.J. Yoon, H. Lee, “Substituent Effect in Amine-CO2 Interaction Investigated by NMR and IR Spectroscopies”, Chem Lett., 32(4) (2003) 324-325. [100] Z.H. Lan, H.A. Mottola, “Carbon Dioxide-Enhanced Luminol Chemiluminescence in the Absence of Added Oxidant”, Analyst, 121(2) (1996), 211-218. [101] M.F. Choi, P. Hawkins, “Novel Dye-Solvent Solutions for the Simultaneous Detection of Oxygen and Carbon Dioxide”, Anal. Chem. 67 (1995) 3897–3902. [102] C.G. Cooney, B.C. Towe, R.C. Eyster, “Optical pH, Oxygen and Carbon Dioxide Monitoring Using a Microdialysis Approach”, Sens. Actuators B, 69 (2000) 183- 188. [1031] B.H. Weigl, A. Holobar, N.V. Rodriguez, O.S. Wolfbeis, “Chemically and Mechanically Resistant Carbon Dioxide Optrode Based on a covalently Immobilised pH Indicator”, Anal. Chim. Acta 282 (1993) 335–343. [104] B.H. Weigl, O.S. Wolfbeis, “New Hydrophobic Materials for Optical Carbon Dioxide Sensors Based on Ion Pairing”, Anal. Chim. Acta 302 (1995) 249–254. [105] J. Dubendorfer, R.E. Kunz, G. Jobst, I. Moser, G. Urban, “Integrated Optical pH Sensor Using Replicated Chirped Grating Coupler Sensor Chips”, Sens. Actuators B, 50 (1998) 210–219. [106] E.M. Hampe, D.M. Rudkevich, “Reversible Covalent Chemistry of CO2”, Chem. Commun., (2002) 1450–1451. [107] A. Mills, Q. Chang, N. McMurray, “Equilibrium Studies on Colorimetric Plastic Film Sensors for Carbon Dioxide”, Anal. Chem. 64 (1992) 1383-1389. [108] B.H. Weigl, O.S. Wolfbeis, “Sensitivity Studies on Optical Carbon Dioxide Sensors Based on Ion Pairing”, Sens. Actuators B, 28 (1995) 151-156.
151
[109] J.W. Parker, O. Laksin, C. Yu, M.L. Lau, S. Klima, R. Fisher, I. Scott, B.W. Atwater, “Fiber Optic Sensors for pH and Carbon Dioxide Using a Self- Referencing Dye”, Anal. Chem., 65 (1993) 2329–2334. [110] A. Mills, C. Chang, “Fluorescent Plastic Thin Film Sensor for Carbon Dioxide”, Analyst, 118(7) (1993) 839-843. [111] N. Nakamura, Y. Amao, “An Optical Sensor for CO2 Using Thymol Blue and Europium (III) Complex Composite Film”, Sens. Actuators B, 92 (2003) 98–101. [112] M. Rocchia, E. Garrone, F. Geobaldo, L. Boarino, M. J. Sailor, “Sensing CO2 in a Chemically Modified Porous Silicon Film”, Phys. Stat. Sol. (a), 197(2) (2003) 365– 369. [113] C. von Bültzingslöwen, A.K. McEvoy, C. McDonagh, B.D. MacCraith, I. Klimant, C. Krause, O.S. Wolfbeis, “Sol–Gel Based Optical Carbon Dioxide Sensor Employing Dual Luminophore Referencing For Application in Food Packaging Technology”, Analyst, 127(11) (2002) 1478–1483. [114] O.S. Wolfbeis, L.J. Weis, M.J. P. Leiner, W.E. Ziegler, “Fiber-Optic Fluorosensor for Oxygen and Carbon Dioxide”, Anal. Chem., 60 (1988) 2028-2030. [115] M.D. Marazuela A, M.C. Moreno-Bondi, G. Orellana, “Enhanced Performance of a Fiber-Optic Luminescence CO2 Sensor Using Carbonic Anhydrase”, Sens. Actuators B, 29 (1995) 126-131. [116] O. Fatibello-Filho, J. F. de Andrade, A.A. Suleiman, G.G. Guilbault, “Piezoelectric Crystal Monitor for Carbon Dioxide in Fermentation Processes”, Anal. Chem., 61 (1989) 746-748. [117] R. Zhou, S. Vaihinger, K.E. Geckeler, W. Gopel, “Reliable CO2 Sensors with Silicon Based Polymers on Quartz Microbalance Transducers”, Sens. Actuators B, 18-19 (1994) 415-420. [118] M.T. Gomes, A.C. Duarte, J.P. Oliveira, “Comparison of Two Methods for Coating Piezoelectric Crystals”, Anal. Chim. Acta, 300 (1995) 329-334. [119] M.T. Gomes, A.C. Duarte, J.P. Oliveira, “Detection of CO2 Using Quarts Crystal Microbalance”, Sens. Actuators B, 26-27 (1995) 191-194. [120] M. T. Gomes, A.C. Duarte, J. P. Oliveira, “The Utilization of a Piezoelectric Quartz Crystal for Measuring Carbon Dioxide in Wine ” Anal. Chim. Acta, 327 (1996) 95-100.
152
[121] M.T.S.R. Games, T.A. Rocha, A.C. Duarte, J.A.B.P. Oliveira, “The Performance of a Tetramethylammonium Fluoride Tetrahydrate Coated Piezoelectric Crystal for Carbon Dioxide Detection”, Anal. Chim. Acta, 335 (1996) 235–238. [122] M.T.S.R. Gomes, T.A.P. Rocha, A.C. Duarte, J.A.B.P. Oliveira, “Quantification of CO2 in Wines with Piezoelectric Crystals Coated with Tetramethylammonium Fluoride and Comparison with Other Methods”, Analusis, 26 (1998) 179-182. [123] M.T.S.R. Gomes, P.S.T. Nogueira, J.B.P. Oliveira, “Quantification of CO2, SO2, NH3, and H2S with a Single Coated Piezoelectric Quartz Crystal”, Sens. Actuators B, 68 (2000) 218–222. [124] M.S. Nieuwenhuizen, A.J. Nederlof, “A SAW Gas Sensor for Carbon Dioxide and Water. Preliminary Experiments”, Sens. Actuators B, 2 (1990) 97-101. [125] P. C. H. Li, M. Thompson, “Potential Carbon Dioxide Sensing Based on Recognition by Trans-(Carbonyl hydroxyl bis-(Triphenyl phosphine) Rhodium(I)) Deposited on Acoustic Wave Devices”, Analyst, 119(9) (1994), 1947-1951. [126] P.C.H. Li, M. Thompson, “Carbon Dioxide Recognition by Rhodium (I) Complexes Studied by Infrared and Nuclear Magnetic Resonance Spectroscopies and Acoustic Wave Sensor”, Analyst, 120(10) (1995) 2529-2535. [127] K. Korsah, C.L. Ma, B. Dress, “Harmonic Frequency Analysis of SAW Resonator Chemical Sensors: Application to the Detection of Carbon Dioxide and Humidity”, Sens. Actuators B, 50 (1998) 110-116. [128] A.E. Hoyt, A.J. Ricco, J.W. Bartholomew and G.C. Osbourn, “SAW Sensors for the Room Temperature Measurement of CO2 and Relative Humidity”, Anal. Chem., 70 (1998) 2137-2145. [129] J.W. Grate, M. Klusty, R.A. McGill, M.H. Abraham, G. Whiting, J. Adonian- Haftvan, “The Predominant Role of Swelling-Induced Modulus Change of the Sorbent Phase in Determining the Responses of Polymer-Coated Surface Acoustic Wave Vapor Sensors”, Anal. Chem., 64 (1992) 610-624. [130] J. Lin, M. Heurich, V. Schlichting, E. Obermeier, “Characterization and Optimization of a CO2 Sensitive Organically Modified Silicate with Respect to Its Use as a Gas Sensor”, Sens. Actuators B, 13-14 (1993) 528-529. [131] C.K. Chiang, C.R. Fincher Jr., Y.W. Park, A.J. Heeger, H. Shirakawa, E.J. Louis, “Electrical Conductivity in Doped Polyacethylene” Phys. Rev. Lett., 39 (1977) 1098-1101.
153
[132] C.M. Mikulski, P. J. Russo, M. S. Saran, A.G. MacDiarmid, A.F. Garito, A.J. Heeger, “Synthesis and Structure of Metallic Polymeric Sulfur Nitride, (SN)x and Its Precursor, Disulfur Dinitride, S2N2”, J. Am. Chem. Soc. 97(22) (1975), 6358- 6363. [133] C.M. Mikulski, A.G. MacDiarmid, A.F. Garito, A.J. Heeger, “Stability of Polymeric Sulfur Nitride, (SN)x to Air, Oxygen, and Water Vapor”, Inorg. Chem., 15(11) (1976) 2943-2945. [134] R. D. Smith, J. R. Wyatt, J.J. DeCorpo, F.E. Saalfeld, M.J. Moran, A.G. MacDiarmid, “A Study of the Vapor Phase Species of Polymeric Sulfur Nitride, (SN)x”, J. Am. Chem. Soc. 99(6) (1977) 1726-1730. [135] W. P. Su, J. R. Schrieffer, A. J. Heeger, “Solitons in Polyacethylene”, Phys. Rev. Lett. 42 (1979) 1698-1701. [136] W. P. Su, J. R. Schrieffer, A. J. Heeger, “Soliton Excitations in Polyacetylene”, Phys. Rev. B, 22 (1980) 2099-2111. [137] A.J. Heeger “Semiconducting and Metallic Polymers: the Fourth Generation of Polymeric Materials”, Synth. Metals, 125 (2002) 23-42. [138] A. J. Heeger, “Charge Transfer in Conductive Polymers”, Faraday Discuss. Chem Soc., 88 (1989) 203-211. [139] T.J. Lewis, “A Simple General Method for Charge Transfer in Polymers”, Faraday Discuss. Chem Soc., 88 (1989) 189-201. [140] G. Wegner, J. Ruhe, “The Structural Background of Charge-Carrier Motion in Conducting Polymers”, Faraday Discuss. Chem Soc., 88 (1989) 333-349. [141] S. Roth, H. Bleier, W. Pukacki, “Charge Transport in Conducting Polymers”, Faraday Discuss. Chem Soc., 88 (1989) 223-233. [142] A.J. Heeger, “Solitons in Conducting Polymers”, Rev. Modern Phys. 60(3) (1989) 781-851. [143] C.B. Duke, E.M. Conwell, A. Paton, “Localized Molecular Excitons in Polyaniline”, Chem. Phys. Lett., 131(1-2) (1986) 82-86. [144] R.R. Chance, D.S. Boudreaux, J.F. Wolf, L.W. Shacklette, R. Silbey, B. Themas, J.M. Andre, J.L. Bredas, “Polyaniline: A Theoretical Study”, Synthetic Metals, 15 (1986), 105-114.
154
[145] F. Zuo, M. Angelopoulos, A.G. MacDiarmid, A.J. Epstein, “Transport Studies of Protonated Emeraldine Polymer: A Granular Polymeric Metal System”, Phys. Rev. B, 36(6) (1987) 3475-3478. [146] A. J. Epstein, A.G. MacDiarmid, “Polaron and Bipolaron defects in Polymers: Polyaniline”, J. Molec. Electronics, 4 (1988) 161-165. [147] R.P. McCall, J.M. Ginder, M.G. Roe, G.E. Asturias, E.M. Scherr, A.G. MacDiarmid, A.J. Epstein, “Massive Polarons in Large-energy-gap Polymers”, Phys. Rev. B, 39(14) (1989) 10174-10178. [148] T. Oshawa, T. Kabata, O. Kimura, K. Yoshino, “Polaronic Transitions in Electrochemical Polymerized Polyaniline”, Synthetic Metals, 29 (1989) E203- E210. [149] Y.H. Kim, C. Foster, J. Chiang, A.J. Heeger, “Localized Charge Excitations in Polyaniline: Infrared Photoexcitation and Protonation Studies”, Synthetic Metals, 29 (1989) E285-E290. [150] H.H.S. Javadi, F. Zuo, K.R. Cromack, M. Angelopoulos, A.G. MacDiarmid, A.J. Epstein, “Charge Transport in the “Emeraldine” Form of Polyaniline”, Synthetic Metals 29 (1989) E409-E416 [151] Y.H. Kim, S.D. Phillips, M.J. Nowak, D. Spiegel, C.M. Foster, G. Yu, J.C. Chiang, A.J. Heeger, “Localization of Charge Excitations in Polyaniline”, Synthetic Metals, 29 (1989) E291-E296. [152] J.K. Avlyanov, S.A. Yakimov, T.A. Usmanov, “Kinetics and Deprotonation of Polyanilines”, Synthetic Metals, 41-43 (1991) 635-639. [153] Z.H. Wang, C. Li, E.M. Scherr, A.G. MacDiarmid, A.J. Epstein, “Three Dimensionality of “Metallic” States in Conducting Polymers: Polyaniline”, Phys. Rev. Lett. 66(13) (1991) 1745-1748. [154] J.G. Masters, Y. Sun, A.G. MacDiarmid, A.J. Epstein, “Polyaniline: Allowed Oxidation States”, Synthetic Metals, 41-43 (1991) 715-718. [155] Z.H. Wang, E.M. Scherr, A.G. MacDiarmid, A.J. Epstein, “Transport and EPR Studies of Polyaniline: a Quasi-on-dimensional Conductor with Three- dimensional “Metallic” States”, Phys. Rev. B, 45(8) (1992), 4190-4202. [156] E.T. Kang, K.G. Neoh, K.L. Tan, “The Protonation-Deprotonation Hysteresis in Polyaniline”, Polymer, 37(6) (1996) 925-929.
155
[157] E.T. Kang, K.G. Neoh, K.L. Tan, “Protonation and Deprotonation of Polyaniline Films and Powders Revisited”, Synthetic Metals 68 (1995) 141-144. [158] E.T. Kang, Z.F. Li, K.G. Neoh, Y.Q. Dong, K.L. Tan, “Protonation and Deprotonation of Polyaniline Films and Powders: Effects of Acid and Base Concentrations on the Surface Intrinsic Oxidation States”, Synthetic Metals, 92 (1998) 167-171. [159] B. Wessling, H. Volk, “Conversion of Doped Polyaniline from the Amorphous to a Partially Crystalline State”, Synthetic Metals, 16 (1986) 127-131. [160] J.C. Chiang, A.G. MacDiarmid, “Polyaniline: Protonic Acid Doping of the Emeraldine Form to the Metallic Regime”, Synthetic Metals, 13 (1986) 193-205. [161] A.C. MacDiarmid, J.C. Chiang, A.F. Richter, A.J. Epstein, “Polyaniline: A New Concept in Conducting Polymers”, Synthetic Metals, 18 (1987) 285-290. [162] W.S. Huang, B.H. Humphrey, A.G. MacDiarmid, “Polyaniline, a Novel Conducting Polymer”, J. Chem. Soc. Faraday Trans 1, 82 (1986) 2385-2400. [163] A. J. Epstein, J.M. Ginder, F. Zuo, H.S. Woo, D.B. Tanner, A.F. Richter, M. Angelopoulos, W.S. Huang, A.G. MacDiarmid, “Insulator-to-Metal Transition in Polyaniline: Effect of Protonation in Emeraldine”, Synthetic Metals, 21 (1987) 63-70. [164] W.S. Huang, A.G. MacDiarmid, A.J. Epstein, “Polyaniline: Non-Oxidative Doping of the Emeraldine Base to the Metallic Regime”, J. Chem. Soc., Chem. Commun., (1987) 1784-1786. [165] A.J. Epstein, A.G. MacDiarmid, “Protonation of Emeraldine: Formation of a Granular Polaronic Polymer Metal”, Mol. Cryst. Liq. Cryst., 160 (1988) 165-173. [166] A. Ray, A.F. Richter, A.G. MacDiarmid, A.J. Epstein, “Polyaniline: Protonation/Deprotonation of Amine and Imine Sites”, Synthetic Metals, 29 (1989) E151-E156. [167] M. Angelopoulos, G.E. Asturias, S.P. Ermer, A. Ray, E.M. Scherr, A.G. MacDiarmid, “Polyaniline: Solution, Films and Oxidation States”, Mol. Cryst. Liq. Cryst. 160 (1988) 151-163. [168] G. E. Asturias, A.G. MacDiarmid, R.P. McCall, A.J. Epstein, “The Oxidation State of “Emeraldine” Base”, Synthetic Metals, 29 (1989) E157-E162. [169] S. Li, H. Dong, Y. Cao, “Synthesis and Characterization of Soluble Polyaniline”, Synthetic Metals, 29 (1989) E329-E336.
156
[170] R.P. McCall, J.M. Ginder, J.M. Leng, H.Y. Ye, S.K. Manohar, J.G. Masters, J.E. Asturias, A.G. MacDiarmid, “Spectroscopy and Defect States in Polyaniline”, Phys. Rev. B, 41(8) (1990) 5202-5213. [171] A. Ray, G.E. Asturias, D.L. Kershner, A.F. Richter, A.G. MacDiarmid, A.J. Epstein, “Polyaniline: Doping, Structure and Derivatives”, Synthetic Metals, 29 (1989) E141-E150. [172] W.S. Huang, A.G. MacDiarmid, “Optical Properties of Polyaniline”, Polymer, 34(9) (1993) 1833-1845. [173] J.E. de Albuquerque, L.H.C. Mattoso, D.T. Balogh, R.M. Faria, J.G. Masters, A.G. MacDiarmid, “A Simple Method to Estimate the Oxidation State of Polyanilines”, Synthetic Metals, 113 (2000) 19-22. [174] J.E. de Albuquerque, L.H.C. Mattoso, R.M. Faria, J.G. Masters, A.G. MacDiarmid, “Study of the Interconversion of Polyaniline Oxidation States by Optical Absorption Spectroscopy”, Synthetic Metals 146 (2004) 1-10. [175] K. Ogura, H. Shiigi, “A CO2 Sensing Composite Film Consisting of Base-Type Polyaniline and Poly (Vinyl Alcohol)”, Electrochem. Solid State Lett., 2(9) (1999) 478-480. [176] K. Ogura, H. Shiigi, M. Nakayama, A. Ogawa, “Thermal Properties of Poly(anthranilic acid) (PANA) and Humidity-Sensitive Composites Derived from Heat-Treated PANA and Poly (Vinyl Alcohol)”, J. Polym. Sci Part A: Polym. Chem., 37 (1999) 4458-4465. [177] K. Ogura, H. Shiigi, T. Oho, T. Tonosaki, “A CO2 Sensor with Polymer Composite Operating at Ordinary Temperature”, J. Electrochem. Soc, 147(11) (2000) 4351-4355. [178] H. Shiigi, T. Oho, T. Tonosaki, K. Ogura, “CO2 Sensitive Characteristics of Base- type Polyaniline / Poly (Vinyl Alcohol) Composites at Room Temperature and Effects of Coexisting Gases”, Electrochemistry, 65(12) (2001) 997-1001. [179] T. Tonosaki, T. Oho, H. Shiigi, K. Isomura, K. Ogura, “Highly Sensitive CO2 Sensor with Polymer Composite Operating at Room Temperature”, Analytical Sciences 17(supplement) (2001) i249-i252. [180] T. Oho, T. Tonosaki, K. Isomura, K. Ogura, “A CO2 Sensor Operating Under High Humidity” J. Electroanal. Chem., 522 (2002) 173-178.
157
[181] F. Koleli, T. Ropke, C.H. Hamann, “The Reduction of CO2 on Polyaniline Electrode in a Membrane Cell”, Synthetic Metals, 140 (2004) 65-68. [182] A.G. MacDiarmid, A.J. Epstein, “Polyanilines: A Novel Class of Conducting Polymers”, Chem, Soc. Faraday Discuss., 88 (1989) 317-332. [183] A.G. MacDiarmid, G.E. Asturias, D.L. Kershner, S.K. Manohar, A. Ray, E.M. Scherr, Y. Sun, X. Tang, A.J. Epstein, “The Polyanilines: Processing, Molecular Weight, Oxidation States and Derivatives”, Polym. Prepr., 30(1) (1989) 147-148. [184] E.M. Scherr, A.G. MacDiarmid, S.K. Manohar, J.G. Masters, Y. Sun, X. Tang, M.A. Druy, P.J. Glatkowski, V.B. Cajipe, J.E. Fischer, K.R. Cromack, M.E. Jozefowicz, J.M. Ginder, R.P. McCall, A.J. Epstein, “Polyaniline: Oriented Films and Fibers”, Synthetic Metals, 41-43 (1991) 735-738. [185] A.G. Green, A.E. Woodhead, “Aniline Black and Allied Compounds-Part 1”, J. Chem. Soc., 97 (1910) 2388-2403. [186] E.W. Paul, A.J. Ricco, M.S. Wrighton, “Resistance of Polyaniline Films as a Function of Electrochemical Potential and the Fabrication of Polyaniline-based Microelectronic Devices”, J. Phys. Chem., 89 (1985) 1441-1447. [187] Y. Wei, G.-W. Jang, K.F. Hsueh, E.M. Scherr, A.G. MacDiarmid, A.J. Epstein, “Thermal Transitions and Mechanical Properties of Films of Chemically Prepared Polyaniline”, Polymer, 33(2) (1992) 314-322. [188] A.G. MacDiarmid, A.J. Epstein, “The Concept of Secondary Doping as Applied to Polyaniline”, Synthetic Metals, 65 (1994), 103-116. [189] A.G. MacDiarmid, A.J. Epstein, “Secondary Doping in Polyaniline”, Synthetic Metals, 69 (1995) 85-92. [190] W. Zheng, A.G. MacDiarmid, “Conformational Effects in Doped Polyaniline: Protonation of Amine and Imine Sites vs. Protonation of Only the Imine Sites”, Polym. Prepr., 36(2) (1995) 73-74. [191] J.P. Travers. M. Nechtschein, “Water Effects in Polyaniline; A New Conduction Process”, Synthetic Metals, 21 (1987) 135-141. [192] M. Nechtschein, C. Santier, J.P. Travers, J. Chroboczek, A. Alix, M. Ripert, “Water Effects on Polyaniline: NMR and Transport Studies”, Synthetic Metals, 18 (1987) 311-316. [193] A. Alix, V. Lemoine, M. Nechtschein, J.P. Travers, C. Menardo, “Water Absorption Study in Polayniline”, Synthetic Metals 29 (1989) E457-E462.
158
[194] M. Angelopoulos, A. Ray, A.G. MacDiarmid, A.J. Epstein, “Polyaniline: Processing from Aqueous Solutions and Effect of Water Vapor on Conductivity”, Synthetic Metals, 21 (1987) 21-30. [195] S.P. Armes, M. Aldissi, S. Agnew, S. Gottesfeld, “Synthesis and Characterization of Aqueous Colloidal Dispersions of Poly (Vinyl Alcohol) / Polyaniline Particles”, Mol. Cryst. Liq. Cryst., 190 (1990) 63-74. [196] K. Ogura, M. Kokura, M. Nakayama, “A Conductive and Humidity Sensitive Composite Film Derived from Poly (o-Phenilenediamine) and Poly (Vinyl Alcohol)”, J. Electrochem. Soc., 142(9) (1995) 152-153. [197] K. Ogura, H. Shiigi, M. Nakayama, “A New Humidity Sensor Using Composite Film Derived from Poly (o-Phenilenediamine) and Poly (Vinyl Alcohol)”, J. Electrochem. Soc., 143(9) (1996) 2925-2930. [198] K. Ogura, H. Shiigi, M. Nakayama, A. Fujii, “Thermogravimetric/Mass and Infrared Spectroscopic Properties and Humidity Sensitivity of Polyaniline Derivatives / Poly (Vinyl Alcohol) Composites”, J. Electrochem. Soc., 145(10) (1998) 3351-3357. [199] K. Ogura, T. Saino, M. Nakayama, H. Shiigi, “The Humidity Dependence of the Electrical Conductivity of a Soluble Polyaniline-Poly (Vinyl Alcohol) Composite Film”, J. Mater Chem., 7(12) (1997) 2363-2366. [200] S.A. Chen, W.G. Fang, “Electrical Conductive Polyaniline-Poly (Vinyl Alcohol) Composite Films: Physical Properties and Morphological Structures”, Macromolecules, 24 (1991) 1242-1248. [201] T. Nagaoka, H. Nakao, T. Suyama, K. Ogura, “Polyaniline-Poly (Vinyl Alcohol) Dispersions for Controlled Ion Exchange of Organic Sulfonates”, Analyst, 122 (1997) 1399-1404. [202] H. Nakao, T. Nagaoka, K. Ogura, “Ion-Exchange Ability of Polyaniline-Poly (Vinyl Alcohol) Colloids with Various Anions”, Analytical Sciences, 13 (1997) 327-331 [203] J. Stejskal, P. Kratochvil, “Polayniline Dispersions 5. Poly (Vinyl Alcohol) and Poly (N-Vinyl Pyrrolidone) as Steric Stabilizers”, Langmuir, 12 (1996) 3389- 3392. [204] J. Stejskal, M. Spirkova, P. Kratochvil, “Polyaniline Dispersions 4. Polymerization Seeded by Polyaniline Particles”, Acta Polymer., 45 (1994) 385- 388.
159
[205] Y. Cao, P. Smith, A.J. Heeger, “Counter-ion Induced Processibility of Conducting Polyaniline”, Synthetic Metals, 55-57 (1993) 3514-3519. [206] M.E. Jozefowicz. , R. Laversanne, H.H.S. Javadi, A.J. Epstein, J.P. Pouget, X. Tang, A.G. MacDiarmid, “Multiple Lattice Phases and Polaron-Lattice-Spinless- Defect Competitions in Polyaniline”, Phys. Rev. B, 39(17) (1989) 12958-12961. [207] H.H.S. Javadi, M. Angelopoulos, A.G. MacDiarmid, A.J. Epstein, “Conduction Mechanism of Polyaniline: Effect of Moisture”, Synthetic Metals, 26 (1988)1-8. [208] F. Zuo, M. Angelopoulos, A.G. MacDiarmid, A.J. Epstein, “A.C. Conductivity of Emeraldine Polymer ”, Phys. Rev B, 39(6) (1989) 3570-3578. [209] J. R. Macdonald, W.B. Johnson, in: J. Ross Macdonald (Ed.), Impedance Spectroscopy Emphasizing Solid Materials and Systems, John Wiley and Sons, New York, 1987, pp. 16. [210] A. Revzin, R.G. Tompkins, M. Toner, “Surface Engineering with Poly(ethylene glycol) Photolithography to Create High-Density Cell Arrays on Glass”, Langmuir, 19 (2002) 9855-9862. [211] M.A.M.J. van Zandvoort, D.L.J. Vossen, G. van Ginkel, R. Torre, P. Bartolini, M. Ricci, J. Thomas-Oates, H. Zuilhof, “Spectral Characterization of Fluorescent 5- Iodoacetamidotetramethylrhodamine and its N-Acetylcysteine Derivative”, Phys. Chem. Chem. Phys., 1 (1999) 4571-4582. [212] W. Zheng, M. Angelopoulos, A.J. Epstein, A.G. MacDiarmid, “Concentration Dependence of Aggregation of Polyaniline in NMP Solutions of Resulting Cast Films”, Macromolecules, 30 (1997) 7634-7637. [213] J. Tang, X. Jing, B. Wang, F. Wang, “Infrared Spectra of Soluble Polyaniline”, Synthetic Metals 24 (1988) 231-238. [214] L. Ding, X. Wang, R.V. Gregory, “Thermal Properties of Chemical Synthesized Polyaniline (EB) Powder”, Synthetic Metals, 104 (1999) 73-78. [215] S.A. Chen, H.T. Lee, “Polyaniline Plasticized with 1-Methyl–2-Pyrrolidone: Structure and Doping Behavior”, Macromolecules, 26 (1993) 3254-3261. [216] W. Zheng, M. Angelopoulos, A.J. Epstein, A.G. MacDiarmid, “Experimental Evidence for Hydrogen Bonding in Polyaniline: Mechanism of Aggregate Formation and Dependency on Oxidation State”, Macromolecules, 30 (1997) 2953-2955.
160
[217] M. Angelopoulos, R. Dipietro, W.G. Zheng, A.G. MacDiarmid, “Effect of Selected Processing Parameters on Solution Properties and Morphology of Polyaniline and Impact on Conductivity”, Synthetic Metals, 84 (1997) 35-39. [218] R. Mathew, B.R. Mattes, M.P. Espe, “A Solid NMR Characterization of Cross- Linked Polyaniline Powder”, Synthetic Metals, 131 (2002) 141-147. [219] R. Mathew, D. Yang, B.R. Mattes, M.P. Espe, “Effect of Elevated Temperature on the Reactivity and Structure of Polyaniline”, Macromolecules, 35 (2002) 7575- 7581. [220] E.J. Oh, Y. Min, J.M. Weisinger, S.K. Manohar, E.M. Scherr, P.J. Prest, A.G. MacDiarmid, A.J. Epstein, “Polyaniline: Dependency of Selected Properties on Molecular Weight”, Synthetic Metals, 55-57 (1993) 977-982. [221] A.J. Milton, A.P. Monkman, “A Comparative Study of Polyaniline Films Using Thermal Analyses and IR Spectroscopy”, J. Phys. D, Appl. Phys. 26 (1993) 1468- 1474. [222] J.P. Pouget, M.E. Jozefowicz, A.J. Epstein, X. Tang, A.G. MacDiarmid, “X-ray Structure of Polyaniline”, Macromolecules, 24 (1991) 779-789. [223] J.P. Pouget, M. Laridjiani, M.E. Jozefowicz, A.J. Epstein, E.M. Scherr, A.G. MacDiarmid, “Structural Aspects of the Polyaniline Family of Electronic Polymers”, Synthetic Metals, 51 (1992) 95-101. [224] M.E. Jozefowicz, A.J. Epstein, J.P. Pouget, J.G. Masters, A. Ray, Y. Sun, X. Tang, A.G. MacDiarmid, “X-Ray Structure of Polyanilines”, Synthetic Metals, 41-43 (1991) 723-726. [225] L.F. Warren, J.A. Walker, D.P. Anderson, C.G. Rhodes, L.J. Buckley, “A Study of Conducting Polymer Morphology”, J. Electrochem. Soc., 136 (1989) 2286- 2295. [226] S. M. Long, K.R. Brennan, A. Saprigin, R.S. Kohlman, A.J. Epstein, M. Angelopoulos, S.L. Buchwalter, A. Rossi, W. Zheng, A.G. MacDiarmid, “Aggregation and Interchain “Self” Doping in Emeraldine Base”, Synthetic Metals, 84 (1997) 809-810. [227] J.Y. Shimano, A.G. MacDiarmid, “Phase Segregation in Polyaniline: A Dynamic Block Copolymer”, Synthetic Metals, 119 (2001) 365-366. [228] J.Y. Shimano, A.G. MacDiarmid, “Polyaniline, a Dynamic Block Copolymer: Key to Attaining its Intrinsic Conductivity?”, Synthetic Metals, 123 (2001) 251- 262.
161
[229] Y.M. Lee, J.H. Kim, J.S. Kang, S.Y. Ha, “Annealing Effects of Dilute Polyaniline/NMP Solution”, Macromolecules, 33 (2000) 7431-7439. [230] M.E. Jozefowicz, A.J. Epstein, J.P. Pouget, J.G. Masters, A. Ray, A.G. MacDiarmid, “X-Ray Structure of the Polyaniline Derivative Poly(o-toluidine): The Structural Origin of Charge Localization”, Macromolecules., 24 (1991) 5863- 5866. [231] S.A. Chen, K.-R. Chuang, C.-I. Chao, H.-T. Lee, “White-Light Emission from Electroluminescence Diode with Polyaniline as the Emitting Layer”, Synthetic Metals, 82(3) (1996) 207-210. [232] M. Wan, J.J. Yang, “Mechanism of Proton Doping in Polyaniline”, Appl. Polym. Sci., 55 (1995) 399-405. [233] A.J. Epstein, A.G. MacDiarmid “Novel Concepts in Electronic Polymers: Polyaniline and its Derivatives”, Makromol. Chem, Macromol. Symp., 51 (1991) 217-234. [234] M.E. Jozefowicz, A.J. Epstein, X. Tang, “Protonic Acid Doping of Two Classes of the Emeraldine Form of Polyaniline”, Synthetic Metals, 46 (1992) 336-337. [235] H. Reiss, “Theoretical Analysis of Protonic Acid Doping of the Emeraldine Form of Polayniline”, J. Phys. Chem., 92 (1988) 3657-3662. [236] H. Reiss, “Note on the Theory of Protonic Acid Doping of Polyaniline”, Synthetic Metals, 30 (1989) 257-263. [237] J. Hammerschmidt, “Studies of Polyvinyl Alcohol under Temperature and Humidity Control”, University of Minnesota. [238] E. Creutz, R.R. Wilson, “Absorption of Water by Films of Cellophane and Polyvinyl Alcohol”, J. Chem. Phys., 14(12) (1946) 725-728. [239] J. Tang, X. Jing, B. Wang, F. Wang, “Infrared Spectra of Soluble Polyaniline”, Synthetic Metals, 24 (1988) 231-238.
162
6. APPENDIX
4.08
4.1
4.12
4.14
4.16
4.18
4.2
4.22
0 1500 3000 4500 6000 7500 9000 10500 12000 13500
Time (min)
Log R
Ar + 5% CO
2
Ar + 5% CO
2
Ar Ar Ar Ar Ar + 5% CO2
Ar + 5% CO
2
Ar + 5% CO
2
Figure 21a. Time variation of the resistance of the composite film in Ar and Ar + 5%
CO2 in a fixed humidity level given by the supersaturated solution of NaCl in water (RH ≅ 60%)-cycle 2 in table 4.
163
4.11
4.12
4.13
4.14
4.15
4.16
4.17
4.18
4.19
4.2
4.21
0 2000 4000 6000 8000 10000 12000 14000 16000
Time (min)
Log R
Argon
Ar + 5% CO2
Figure 22a. Trend of the resistance response of the composite film in Ar and Ar + 5%
CO2 at a fixed humidity level given by the supersaturated solution of NaCl in water RH ≅ 60%-cycle 2 in table 4.
164
4.15
4.16
4.17
4.18
4.19
4.2
4.21
4.22
4.23
4.24
0 1500 3000 4500 6000 7500 9000 10500 12000 13500
Time (min)
Log R
CO2 CO2 CO2ArAr Ar
Figure 21b. Time variation of the resistance of the composite film in Ar and Ar + 5%
CO2 in a fixed humidity level given by the supersaturated solution of MgCl2 in water (RH
≅ 30%)-cycle 3 in table 4.
165
4.17
4.18
4.19
4.2
4.21
4.22
4.23
4.24
0 2000 4000 6000 8000 10000 12000 14000
Time (Min)
Log R
Argon
Ar + 5% CO2
Figure 22b. Trend of the resistance response of the composite film in Ar and Ar + 5%
CO2 at a fixed humidity level given by the supersaturated solution of MgCl2 in water
(RH ≅ 30%)-cycle 3 in table 4.
166
4.15
4.2
4.25
4.3
4.35
0 2500 5000 7500 10000 12500 15000 17500 20000 22500 25000
Time (min)
Log R
Ar Ar Ar Ar + 5% CO2Ar + 5% CO2
Ar + 5% CO
2
Ar + 5% CO
2
Figure 21c. Time variation of the resistance of the composite film in Ar and Ar + 5%
CO2 in a fixed humidity level given by the supersaturated solution of LiCl in water (RH ≅ 4.5%)-cycle 4 in table 4.
167
4.28
4.29
4.3
4.31
4.32
4.33
4.34
0 5000 10000 15000 20000 25000Time (min)
Log R
Argon
Ar + 5% CO2
Figure 22c. Trend of the resistance response of the composite film in Ar and Ar + 5%
CO2 at a fixed humidity level given by the supersaturated solution of LiCl in water (RH
≅ 4.5%)-cycle 4 in table 4.
168
4.10
4.12
4.14
4.15
4.17
4.19
4.21
4.23
4.24
4.26
4.28
4.30
0 2000 4000 6000 8000 10000 12000 14000 16000
Time (min)
Log R
CO2 CO2Ar Ar
Figure 21d. Time variation of the resistance of the composite film in Ar and Ar + 5%
CO2 in a fixed humidity level given by the supersaturated solution of LiBr in water (RH ≅ 1.5%)-cycle 5 in table 4.
169
4.15
4.17
4.19
4.21
4.23
4.25
4.27
0 2000 4000 6000 8000 10000 12000 14000 16000 18000
Time (min)
Log R
Argon
Ar + 5% CO2
Figure 22d. Trend of the resistance response of the composite film in Ar and Ar + 5%
CO2 at a fixed humidity level given by the supersaturated solution of LiBr in water (RH
≅ 1.5%)-cycle 5 in table 4.
170
4.1
4.11
4.12
4.13
4.14
4.15
4.16
4.17
4.18
0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000
Time (min)
Log R
CO2 CO2 CO2ArAr Ar
Figure 21e. Time variation of the resistance of the composite film in Ar and Ar + 5%
CO2 in a fixed humidity level given by the supersaturated solution of MgCl2 in water (RH
≅ 30%)-cycle 6 in table 4.
171
4.11
4.12
4.13
4.14
4.15
4.16
4.17
4.18
0 5000 10000 15000 20000 25000
Time (min)
Log R
Argon
Ar + 5% CO2
Figure 22e. Time variation of the resistance of the composite film in Ar and Ar + 5%
CO2 in a fixed humidity level given by the supersaturated solution of MgCl2 in water (RH
≅ 30%)-cycle 6 in table 4.
172
4.100
4.120
4.140
4.160
4.180
4.200
4.220
4.240
4.260
4.280
4.300
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 14000 15000 16000
Time (min)
Log R
Ar + 5% CO2 Ar Ar
T90 T90 T90 T90
Ar + 5% CO2
Figure 25a. Example of t90 estimation-cycle 5 in table 4.
173
0
10
20
30
40
50
60
70
80
90
05001000150020002500300035004000
Wavelength (cm-1)
Transm
ittance (a.u.)
3365 cm-1
(free N-H stretching)
3284 cm-1
(H-bonded N-H stretching) 1596 cm-1
(C=C stretching-Quinoid)1497 cm-1
(C=C stretching-Benzenoid)
1299 cm-1
C-N stretching- amine
1145-1164 cm-1
(N=Q=N stretching)
1040 cm-1
(SO2 stretching-
symmetryc in TSA)
Figure 26a. FTIR spectra for an Emeraldine Base treated at 100° C for 1 hour in helium
atmosphere.
174
0
20
40
60
80
100
120
05001000150020002500300035004000
Wavenumber (cm-1)
Transm
ittance (a.u.)
1589 cm-1
(C=C stretching-Quinoid)1485 cm-1
(C=C stretching-Benzenoid)
1296 cm-1
C-N stretching-secondary aromatic amine
1143-1164 cm-1
(N=Q=N stretching)3366 cm-1
(free N-H stretching)
3280 cm-1
(H-bonded N-H stretching)
1039 cm-1
(SO2 stretching-
symmetryc in TSA)
Figure 26b. FTIR spectra for an Emeraldine Base treated at 150° C for 1 hour in helium
atmosphere.
175
0
20
40
60
80
100
120
05001000150020002500300035004000
Wavewnumber (cm-1)
Transm
ittance (a.u.)
3368 cm-1
(free N-H stretching)
3264 cm-1
(H-bonded N-H stretching)
1590 cm-1
(C=C stretching-Quinoid)1496 cm-1
(C=C stretching-Benzenoid)
1289 cm-1
C-N stretching-secondary aromatic amine
1165 cm-1
(N=Q=N stretching)
1037 cm-1
(SO2 stretching-
symmetryc in TSA)
Figure 26c. FTIR spectra for an Emeraldine Base treated at 200° C for 1 hour in helium
atmosphere.
176
0
20
40
60
80
100
120
05001000150020002500300035004000
Wavenumber (cm-1)
Transm
ittance (a.u.)
3365 cm-1
(free N-H stretching)
3284 cm-1
(H-bonded N-H stretching)
1596 cm-1
(C=C stretching-Quinoid)1497 cm-1
(C=C stretching-Benzenoid)
1290 cm-1
C-N stretching-amine
1168 cm-1
(N=Q=N stretching)
1036 cm-1
(SO2 stretching-
symmetryc in TSA)
Figure 26d. FTIR spectra for an Emeraldine Base treated at 250° C for 1 hour in helium
atmosphere.
177
0
20
40
60
80
100
120
140
05001000150020002500300035004000
Wavenumber (cm-1)
Transm
ittance (a.u.)
1170 cm-1
(N=Q=N stretching)
1285 cm-1
C-N stretching-secondary aromatic amine
1499 cm-1
(C=C stretching-Benzenoid)
1594 cm-1
(C=C stretching-Quinoid)
3376 cm-1
(free N-H stretching)
1045 cm-1
(SO2 stretching-
symmetryc in TSA)
Figure 26e. FTIR spectra for an Emeraldine Base treated at 300° C for 1 hour in helium
atmosphere.
178
0
200
400
600
800
1000
1200
1400
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
2 Theta (deg)
Intensity (a.u.)
FWHM: 20.975-18.775=2.2
HW/H: 14.75/25.5=0.57
FWHM
Peak Width
Peak Height
Figure 29a. Method of calculating the FWHM and HW/H for an XRD pattern of
emeraldine base powder heat treated at 100ºC.
179
0.0E+00
1.0E-06
2.0E-06
3.0E-06
4.0E-06
5.0E-06
6.0E-06
0.00 5.00 10.00 15.00 20.00 25.00
Time (s1/2)
Clow frequency (F)
0.0E+00
2.0E-07
4.0E-07
6.0E-07
8.0E-07
1.0E-06
1.2E-06
1.4E-06
1.6E-06
Chigh frequency (F)
Chigh frequency
Clow frequency
pH = 1.75 pH = 1.11
Figure 43a. Capacitance vs. time in pH = 1.75 and pH = 1.11.
180
0
500
1000
1500
2000
2500
01234567pH
Resistance (ohm)
Figure 43b. Total resistance of emeraldine base thin film vs. pH in aqueous HCl
solutions.